Construction of a shuttle expression vector with a promoter functioning in both halophilic Archaea and Bacteria



A shuttle expression vector, designated as pAJ, was constructed based on the Haloferax volcanii-Escherichia coli shuttle vector pSY1. This new construct contains the amyH promoter from Haloarcula hispanica and was able to confer the promoter activity in both Hfx. volcanii and E. coli. pAJ successfully expressed proteins in Hfx. volcanii or E. coli, rendering it feasible to express target proteins in corresponding domains. In addition, pAJ contains a multiple cloning site with 11 restriction sites and a 6×His tag sequence, and the vector size was decreased to 8903 bp. To the best of our knowledge, pAJ is the first reported shuttle expression vector that can express proteins in both Bacteria and Archaea. Importantly, pAJ can even express the haloarchaeal heat shock protein DnaK in both domains. In conclusion, this novel vector only provides researchers with a new means to manipulate genes or express proteins in Haloarchaea but also serves as a convenient tool for the comparative study of the function of some highly conserved genes in Haloarchaea and in Bacteria.


The extremely halophilic Archaea are a group of microorganisms that thrive in hypersaline environments; the family currently possesses 36 recognized genera with 129 species (Oren, 2012). Among them, only limited species, such as Haloferax volcanii, Haloarcula marismortui, and Haloarcula hispanica, can undergo molecular manipulation. While the underdevelopment of gene engineering in Haloarchaea may be ascribed to many reasons, e.g. less selectable markers and limited transformation protocols (Allers & Mevarech, 2005), it is worthwhile to note that the number of available DNA cloning tools for these microorganisms is far less than that for Bacteria or eukaryotes. Therefore, it is necessary to develop more versatile vectors to facilitate the genetic manipulation of these extremely halophilic Archaea.

Compared with other ordinary molecular tools, shuttle vectors are an attractive system because of the ability to propagate in different host species or even diverse genera and domains. The most popular host for these vectors is Escherichia coli. pWL102 (Lam & Doolittle, 1989) is the first shuttle vector between Hfx. volcanii and E. coli. Sequentially, several improved shuttle vectors have been constructed to establish better stability (Holmes et al., 1994) and broader host affinity (Blaseio & Pfeifer, 1990; Cline & Doolittle, 1992). A milestone was the development of pWL204, the first shuttle expression vector with a tRNALys promoter, which rendered it feasible not only to replicate in both host cells, but also to express genes in Hfx. volcanii (Nieuwlandt & Daniels, 1990). To date, however, there is no report of a shuttle vector that confers promoter activity (i.e. drive transcription) of the same promoter in both Bacteria and Haloarchaea. Such a vector will potentially provide a convenient tool to express proteins or comparatively evaluate gene functions in different domains without additional cloning steps.

Previously, we found the promoter region of the halophilic α-amylase (amyH) gene from Har. hispanica could also drive gene expression in E. coli (Zeng et al., 2009). Here, we report the construction of a shuttle expression vector based on the amyH promoter and Hfx. volcanii–E. coli shuttle vector pSY1 (Yang et al., 2003). The new vector, designated as pAJ, has the following three advantages:

  • With a haloarchaeal promoter conferring activity in both Hfx. volcanii and E. coli, pAJ could express proteins in both domains.
  • pAJ harbor a multiple cloning site (MCS) plus a His tag-coding sequence and is therefore a convenient tool for gene manipulation and protein purification.
  • The pAJ vector size (8903 bp) is relatively small compared with either pSY1 (12 209 bp) or many other haloarchaeal vectors, such as pWL102 (10.5 kb; Lam & Doolittle, 1989), pUBP2 (12 kb; Blaseio & Pfeifer, 1990) and pMLH3 (11.3 kb; Holmes et al., 1994), and smaller vectors are more convenient for manipulating large exogenous DNA fragments.

Materials and methods

Strains and culture conditions

The strains and plasmids used in this study are listed in Supporting Information, Table S1. Hfx. volcanii were grown at 42 °C in 18% (w/v) MGM as described by Dyall-Smith (2008). Tranformants are selected, supplemented with 0.3 μg mL−1 novobiocin (Sigma) when necessary. Haloarcula hispanica DSM 4426 was grown in 23% MGM (Dyall-Smith, 2008) and used for cloning the amyH gene. The Natrinema sp. J7-2 strain was grown in 25% (salt, w/v) Halo-2 or in 20% MGM (Mei et al., 2007; Zhang et al., 2009, 2012) and used for cloning the dnaK gene. Escherichia coli were grown at 37 °C in Luria–Bertani (LB) medium (Sambrook et al., 1989), supplemented with ampicillin (final concentration, 100 μg mL−1) when necessary. Primers used for plasmid construction are listed in Table S2.

Construction of pAJ

As shown in Supporting Information, Fig. S1, the construction of pAJ was based on pSY1, a Hfx. volcaniiE. coli shuttle vector carrying a bop promoter (Yang et al., 2003). The amyH promoter (pro amyH) from Har. hispanica was cloned to replace the bop promoter in pSY1, yielding the vector pAB. Because the open reading frame (ORF) of the halophilic β-galactosidase (bgaH) gene (Cline et al., 1989) contains a restriction site (KpnI) incompatible with sequential digestions, the bgaH ORF was replaced by the amyH ORF, yielding pA92-fl1. Next, ApaI, HindIII and PstI in pA92-fl1 were erased by digestion, blunting and ligation to create pA92-fl1D.

To further decrease the vector size, the pBluescript II SK region (pSK) in pA92-fl1 harboring the ColE1 origin and the ampicillin-resistance gene (ampr) was removed at the restriction sites KpnI and XbaI, and then replaced by a modified pUC19 plasmid, pUC19-D. pUC19-D was obtained by removing the two following parts from pUC19: SacI-SspI and SalI-PscI regions, both of which were excised by digestion with the respective restriction enzymes, followed by blunting with S1 nuclease and ligation with DNA ligase. Insertion of pUC19-D into pA92-fl1D at the KpnI/XbaI sites yielded the plasmid pAH, which was 1113 bp smaller than pSY1.

To insert an MCS with more restriction sites in the vector, we constructed two MCS segments by megaprimer PCR. An upstream fragment of the alkaline protease gene (arpE) from Bacillus licheniformis L106 was amplified using a pair of megaprimers, MUarpE-F and MUarpE-R (Table S2). In the megaprimer MUarpE-F, a (CAC)6 tract expressing a hexahistidine (6×His) tag was inserted in the frame at the BclII site after the start codon ATG in the NdeI (CATATG) site (Table S2, Fig. S1). Then the arpE fragment for an alkaline protease gene (arpE) from B. licheniformis L106 was amplified using a pair of megaprimers, MarpE-F and MarpE-R (Table S2). PCR products containing MCS were purified and ligated with pMD19T-simple for sequencing. Two fragments containing MCS were inserted into pAH to give pAHUA. Next, the upstream fragment of arpE was removed from the vector to produce pAHA. Finally, the arpE fragment was removed to generate the plasmid pAJ (Fig. S1). Insertion of the new MCS in the final constructed vector increased the number of usable restriction sites from two (NdeI and NcoI) to 11 (NdeI, BclI, ApaI, AflII, HindIII, BbeI, XhoI, MfeI, PstI, MscI and NcoI), all of which will be used to produce His-tagged proteins except BclI, as BclI digestion will remove the His tag sequence.

Transformation of Hfx. volcanii and E. coli

Transformation of Hfx. volcanii was performed using the polyethylene glycol-mediated spheroplast transformation method (Charlebois et al., 1987; Cline et al., 1989). Transformation of E. coli was carried out following the standard protocol of CaCl2-mediated transformation (Sambrook et al., 1989).

β-Galactosidase assay

The β-galactosidase-specific activity in Hfx. volcanii was analyzed by the ONPG assay as described by Dyall-Smith (2008). The β-galactosidase activity in Hfx. volcanii colonies was visualized by spraying agar plates with X-gal (20 mg mL−1).

Amylase activity assay

The amylase-specific activity in Hfx. volcanii was measured by a method described by Fuwa (1954). The presence of amylase activity on plates was determined as described by Coronado et al. (2000) using 18% MGM supplemented with 2% (w/v) soluble starch. After incubation at 37 °C for about 7 days, plates were flooded with 0.3% I2/0.6% KI solution; a clear zone around the growth indicated starch hydrolysis.

Chloramphenicol acetyltransferase (CAT) assay

The CAT activity in E. coli was quantitatively determined using the CAT ELISA Kit (Roche). Cultures were grown in LB medium to an OD600 nm of 0.3–0.4. Samples were taken, centrifuged, and resuspended in phosphate-buffered saline buffer, prior to cell lysis. After cell extracts were centrifuged to remove intact cells and debris, supernatants were used for CAT and total protein determinations. Protein concentrations were determined using the Bradford assay with Bovine Serum Album (BSA) as standard (Coligan et al., 2003).

WF146 protease activity assay

The proteolytic activity of WF146 protease in E. coli was assayed by the appearance of a hydrolytic zone on 2% (w/v) milk plates, supplemented with 0.1% (w/v) ampicillin. Cultures were grown at 37 °C for 48 h and then cultured at 45 °C to observe the clear zone around the colonies.

Complementation of dnaK-null mutant

Temperature-sensitive mutants, E. coli BM271 and Hfx. volcanii HV66 (both ΔdnaK mutants), were transformed with pAJ-DnaK. The complementation assay of the mutated gene in E. coli BM271 was performed as described by Zhang et al. (2007). For Hfx. volcanii HV66, the method was basically the same as that used for E. coli BM271, except that the strains were grown at 42 °C in 18% MGM supplemented with novobiocin. Successive 10-fold dilutions of the culture samples were dropped onto 18% MGM plates and then incubated for 5–7 days at either 42 or 52 °C. Wild-type strains and ΔdnaK mutant strains containing either no plasmid or empty pAJ were used as controls.

Western blot analysis

Escherichia coli and Hfx. volcanii were grown at 37 °C for 18 h and then at 45 °C for 4 days. Next, 200 μL of each culture was centrifuged at 12 857 g for 10 min, suspended with 1× Sodium dodecyl sulfate (SDS) loading buffer, and incubated at 95 °C for 20 min. An anti-DnaK antibody was prepared by Dr Yuping Huang's group from rabbits (W. Chen, Y. Huang, unpublished data).


Identification of pAJ as a shuttle expression vector to express proteins in halophilic Archaea

To confirm the cross-domain activity of the amyH promoter in pAJ, we first analyzed its potential to express archaeal proteins in Hfx. volcanii. The bgaH gene from Haloferax alicantei (Holmes et al., 1997; Patenge et al., 2000) and the amyH gene from Har. hispanica (Coronado et al., 2000; Zeng et al., 2009) were selected as the reporter genes in this assay. BgaH was cloned into pAJ between HindIII and PstI to generate pAJ-bgaH-HP, which was subsequently transformed into Hfx. volcanii DS52. The X-gal plate assay shows that only strains harboring bgaH-containing plasmids produced blue colonies (Fig. 1b), and the β-galactosidase-specific activity assay also indicates bgaH gene expression (Fig. 1a). In addition, pAJ-bgaH-HP and pAJ-bgaH-XP (insertion between XhoI and PstI) displayed no significant difference, indicating that the gene expression was not affected by the use of different restriction endonuclease sites. Likewise, using the starch-plate assay (Fig. 1d) and amyH-specific activity assay (Fig. 1c), the same conclusion was reached for the amyH gene, which was cloned into pAJ to produce pAJ-amyH-HM (HindIII and MfeI) and pAJ-amyH-XP (XhoI and PstI).

Figure 1.

Reporter gene expression at different insertion sites. (a) The β-galactosidase (BgaH)-specific activity assay in Haloferax volcanii in 25% MGM. HP, insertion between HindIII and PstI. XP, insertion between XhoI and PstI. (b) The β-galactosidase (BgaH) activity assay in Hfx. volcanii on (plate 1) pAJ/DS52, (plate 2) pAB/DS52, (plates 3–5) pAJ-BgaH(HP)/DS52, and (plates 6–8) pAJ-BgaH(XP)/DS52. (c) The amylase (amyH)-specific activity assay in Hfx. volcanii in 25% MGM. HM, insertion between HindIII and MfeI. XP, insertion between XhoI and PstI. (d) The amylase (amyH)-specific activity assay in Hfx. volcanii on (plate 1) pAJ/DS52, (plate 2) pA92-fl1/DS52, (plates 3–5) pAJ-amyH(HM)/DS52, and (plates 6–8) pAJ-amyH(XP)/DS52. (e) The chloramphenicol acetyltransferase (CAT) activity assay in Escherichia coli. AP, insertion between AflII and PstI. MP, insertion between MfeI and PstI. (f) The WF146 protease activity assay in E. coli. (1) pAJ/DS52, (2) pAJ-WF146(NX)/DH5α, and (3) pAJ-WF145(AP)/DH5α.

Identification of pAJ as a shuttle expression vector to express proteins in Bacteria

Next, bacterial proteins were used as indicators to validate the amyH promoter activity in E. coli. The chloramphenicol acetyltransferase (cat) gene from the plasmid pAD123 (Dunn & Handelsman, 1999) was cloned into pAJ to yield pAJ-Cm-AP (AflII and PstI) and pAJ-Cm-MP (MfeI and PstI). The transformants carrying pAJ-Cm-AP or pAJ-Cm-MP were found to display resistance to chloramphenicol (data not shown) and showed similar activities of chloramphenicol acetyltransferase (about 29 μg mg−1) determined using the ELISA method (Fig. 1e). Another reporter gene was the protease WF146 from thermophilic Bacillus sp. WF146 (Wu et al., 2004), which was cloned into pAJ to yield pAJ-WF146-NX (NdeI and XhoI) and pAJ-WF145-AP (ApaI and PstI). Only strains bearing pAJ-WF146-NX and pAJ-WF145-AP formed a hydrolysis zone around the colonies on the milk plate (Fig. 1f), without any significant difference between them.

Complementation of dnaK-null mutants of Hfx. volcanii and E. coli by expressing the haloarchaeal protein DnaK from Natrinema sp. J7-2

DnaK (Hsp70) is a major heat shock protein ubiquitously expressed in prokaryotes; mutations in dnaK usually result in a temperature-sensitive phenotype in the host strains. For instance, dnaK-deletion mutants of E. coli and Hfx. volcanii are inviable or only grow weakly at 42 and 52 °C, respectively, whereas the corresponding wild-type strains grow normally at those temperatures. To examine whether the amyH promoter is able to produce proteins in their active form in both Bacteria and Haloarchaea, dnaK from Haloarchaea Natrinema sp. J7-2 was amplified and cloned into pAJ to yield pAJ-DnaK. pAJ-DnaK and the original vector pAJ were transformed into dnaK-deletion mutants of E. coli BM271 and Hfx. volcanii HV66. In the latter case, pAJ-DnaK was first transformed into E. coli JM110 for demethylation before transformation into Hfx. volcanii HV66. The expression of the target protein in both species was determined by Western blot analysis (Figs 2a and 3a). The activity of DnaK was assessed as described in Materials and methods. As shown in Fig. 2, DnaK from Natrinema sp. J7-2 was able to restore obviously the growth of the E. coli danK mutant strain, despite having no immunological cross-reactivity with the counterpart protein. This result is in agreement with our previous work, in which the protein expression was driven by a bacterial promoter (Zhang et al., 2007). On the other hand, although DnaK from Natrinema sp. J7-2 showed obvious immunological cross-reactivity with that of Hfx. volcanii, it exhibited only a slight complement effect upon the Hfx. volcanii danK mutant strain (Fig. 3). The growth of strain HV66 was always better than that of strain HV66/pAJ or strain HV66/pAJ-DnaK; a possible reason may be that the large vector made the host strain a burden. Interestingly, although strain HV66/pAJ-DanK harbored a larger vector, it exhibited better growth than HV66/pAJ at 52 °C, indicating some exertion of exogenously expressed DnaK (Fig. 3b2). All the experiments were repeated several times. Our results support the conclusion that by employing pAJ, it is possible to express the same protein with an active form in both Hfx. volcanii and E. coli. In addition, the pAJ vector could also be a valuable tool for comparative study of the function of highly conserved genes between Haloarchaea and Bacteria.

Figure 2.

DnaK expression in Bacteria. (a) Determination of the DnaK expression by Western blot. (b) Complementation experiments of the Escherichia coli dnaK-null mutant at 30 °C (b1) and 42 °C (b2). Serial dilutions of bacterial suspensions from 100 to 10−6 were dotted from left to right. J7-2, Natrinema sp. J7-2 (positive control); MC4100, E. coli MC4100 (wild type); ΔK, E. coli dnaK-null mutant strain BM271ΔdnaK; ΔK/pAJ-DnaK, E. coli BM271ΔdnaK carrying pAJ-DnaK; ΔK/pAJ, E. coli BM271ΔdnaK carrying pAJ.

Figure 3.

DnaK expression in Archaea. (a) Determination of DnaK expression by Western blot. (b) Complementation experiments of the Haloferax volcanii danK-null mutant strain at 42 °C (b1) and 52 °C (b2). Serial dilutions of bacterial suspensions from 100 to 10–6 were dotted from left to right. J7-2, Natrinema sp. J7-2 (positive control); H53, Hfx. volcanii H53 (starting strain for HV66); HV66, Hfx. volcanii danK-null mutant strain HV66; HV66/pAJ, HV66 carrying plasmid pAJ; HV66/pAJ-DnaK, HV66 carrying plasmid pAJ-DnaK.


This study deals with the construction and evaluation of a Hfx. volcaniiE. coli shuttle expression vector, designated as pAJ. With the promoter of amyH from Har. hispanica replacing that of haloarchaeal bop (Baliga & DasSarma, 2000; Yang et al., 2003), pAJ is able not only to drive gene transcription in Hfx. volcanii, like the original vector pSY1, but also to express proteins in E. coli. To the best of our knowledge, pAJ is the first documented vector that can express genes in both Haloarchaea and Bacteria under the same promoter. In addition, the vector size was reduced from 12 209 bp (pSY1) to 8903 bp (pAJ), and the number of available restriction enzyme sites was increased from two (NdeI and NcoI) in pSY1 to 11 (NdeI, BclI, ApaI, AflII, HindIII, BbeI, XhoI, MfeI, PstI, MscI, and NcoI) in pAJ. Our results also demonstrate that in both Haloarchaea and Bacteria, the promoter activities of the vector are not affected by where the gene cloning took place in the MCS.

In the case of protein expression in Haloarchaea, the most widely used vector is currently pTA963 from Thorsten Allers (Allers et al., 2010). However, although pTA963 (8340 bp) is a little smaller than pAJ (8903 bp), pTA963 only contains six usable single restriction enzyme sites (BamHI, NdeI, EcoRI, PciI, NcoI, and BspHI), fewer than pAJ. Among the six enzyme sites in pTA963, PciI, NcoI and BspHI are located at the same site with the compatible ends. In addition, only PciI and NdeI could be employed for protein affinity isolation using the His tag; the other two sites, EcoRI and BamHI, would cause gene frameshift mutations if the His tag sequence were used. In contrast, pAJ could provide 10 choices (NdeI, ApaI, AflII, HindIII, BbeI, XhoI, MfeI, PstI, MscI, and NcoI) for expressing 6×His-tagged proteins. Therefore, although the amyH promoter has not yet been compared side-by-side with other widely used haloarchaeal promoters such as fdx from Halobacterium salinarum (Pfeifer et al., 1993; Danner & Soppa, 1996) or tna from Hfx. volcanii (Large et al., 2007), we believe that pAJ should be a good alternative to express proteins in Hfx. volcanii, either for evaluating the gene functions or for collecting exogenously expressed proteins.

Contrary to the well established concept that haloarchaeal proteins typically are not produced in an inactive form in the low-salt cytoplasm of E. coli, our cross-domain complementation assay suggests that at least DnaK from Haloarchaea Natrinema sp. J7-2 could be functionally expressed in Bacteria. A comparative study of evolutionary conserved genes among different species might provide insights into a further understanding of their phylogenies. A point worth emphasizing is that although pAJ may not be an ideal tool for protein production in E. coli, compared with other well developed bacterial expression vectors, it should be a valuable candidate for comparing gene functions between Haloarchaea and Bacteria. That is, cross-domain shuttle expression vectors are not only convenient for gene cloning, but also provide almost identical conditions for such comparative research.

Archaeal promoters have long been believed to resemble eukaryotic counterparts; the amyH promoter from Har. hispanica (Zeng et al., 2009) displays similar core functions to bacterial promoters. The elucidation of the mechanism for this ‘cross-domain activity’ of amyH promoter probably relies on its structural resemblance to bacterial promoters (Yang et al., 2003; Zeng et al., 2009). Studies on promoters conferring the cross-domain activity would provide crucial clues to the evolutionary relationship between basal transcription signals and transcription mechanisms in all three domains of life, which may in turn expand our understanding of the evolution of life (Zeng et al., 2009). Therefore, it is important to identify more promoters like the amyH promoter to add to our knowledge of the evolutionary significance of archaeal promoters. To achieve this, the amyH promoter in our newly constructed vector could be replaced with potential candidate fragments, rendering it a feasible promoter-probing shuttle vector and generating more convenient and useful vectors containing promoters with higher activities in both Bacteria and Archaea. Indeed, we have already isolated several promoters from Natrinema sp. J7-2 (Mei et al., 2007; Zhang et al., 2009, 2012) with the cross-domain activity, some of which exhibit higher activities (J. Lv, X. Chen, unpublished data).


This work was supported by grants from the National Basic Research Program of China (973Program) (2011CB808800), National Natural Science Foundation of China (No. 30970070 and 31000050), and the Chinese 111 project (#B06018). We thank Prof. Bing Tang (Wuhan University, China) for providing plasmid pHSNM. We also thank Dr Dong Chen (Yale University, USA) for critical reading and editing of the manuscript.