Analysis of the Transcriptional Regulator GlpR, Promoter Elements, and Posttranscriptional Processing Involved in Fructose-Induced Activation of the Phosphoenolpyruvate-Dependent Sugar Phosphotransferase System in Haloferax mediterranei

Abstract

Among all known archaeal strains, the phosphoenolpyruvate-dependent phosphotransferase system (PTS) for fructose utilization is used primarily by haloarchaea, which thrive in hypersaline environments, whereas the molecular details of the regulation of the archaeal PTS under fructose induction remain unclear. In this study, we present a comprehensive examination of the regulatory mechanism of the fructose PTS in the haloarchaeon Haloferax mediterranei. With gene knockout and complementation, microarray analysis, and chromatin immunoprecipitation-quantitative PCR (ChIP-qPCR), we revealed that GlpR is the indispensable activator, which specifically binds to the PTS promoter (P_PTS) during fructose induction. Further promoter-scanning mutation indicated that three sites located upstream of the H. mediterranei P_PTS, which are conserved in most haloarchaeal P_PTSs, are involved in this induction. Interestingly, two PTS transcripts (named T₈ and T₁₇) with different lengths of 5′ untranslated region (UTR) were observed, and promoter or 5′ UTR swap experiments indicated that the shorter 5′ UTR was most likely generated from the longer one. Notably, the translation efficiency of the transcript with this shorter 5′ UTR was significantly higher and the ratio of T₈ (with the shorter 5′ UTR) to T₁₇ increased during fructose induction, implying that a posttranscriptional mechanism is also involved in PTS activation. With these insights into the molecular regulation of the haloarchaeal PTS, we have proposed a working model for haloarchaea in response to environmental fructose.

INTRODUCTION

The phosphoenolpyruvate (PEP)-dependent sugar phosphotransferase system (PTS) uses PEP as the phosphoryl donor to phosphorylate sugars for transport into cells (1, 2). A typical PTS contains five proteins, PtsI (or EI), HPr, PtsA, PtsB, and PtsC. Phosphotransfer from PEP to sugar is mediated by these five proteins in a cascade, and the PtsC component at the end of the cascade couples phosphorylation with the translocation of the specific sugars (3, 4). The PTS is an important apparatus for sugar uptake and degradation in bacteria, and most bacteria have been shown to possess at least one complete PTS (4). The PTSs of bacteria can sense the primary metabolic or environmental signal and turn on the uptake system (5, 6). In response to the environmental signal, the derivatives of sugars always serve as positive or negative effectors, while the global or specific transcriptional regulators, in cooperation with cyclic AMP (cAMP) or primary metabolites, are involved in the complicated regulation of the PTS via direct binding to the promoter regions of PTS genes (2, 7).

In contrast to the case for bacteria, research on archaeal PTSs has received attention just in the past few years. The first report that archaea have PTS genes was published in 2006, based on the genome sequencing of the haloarchaeon Haloquadratum walsbyi (8). Recent studies of haloarchaeal genome sequences have indicated that many haloarchaea contain PTS genes (8–11), and 6 out of 24 haloarchaeal genomes have a complete fructose-specific PTS gene cluster, including Haloterrigena turkmenica, Halalkalicoccus jeotgali, Haloarcula hispanica, Haloarcula marismortui, Haloferax volcanii, and Haloferax mediterranei (11). Recently, a functional fructose-specific PTS has been identified in H. volcanii using genetic methods, and fructose was shown to be able to upregulate the transcription of this PTS gene cluster (12), but the molecular details of the fructose-induced PTS activation in archaea remain unclear. Interestingly, earlier research on H. volcanii indicates that a DeoR family transcriptional regulator, GlpR, represses the expression of fructose and glucose metabolic enzymes (2-keto-3-deoxy-d-gluconate kinase [KDGK] and phosphofructokinase [PFK]) at the transcriptional level when cells are grown on glycerol (13). In addition, it was reported that glpR is cotranscribed with the downstream phosphofructokinase gene (fruK) (13). The PTS gene cluster is located just adjacent to glpR-fruK. As an important regulator which usually functions in sugar metabolism in bacteria (14, 15), GlpR is probably involved in the transcriptional regulation of the haloarchaeal PTS cluster. However, as far as we know, the relationship between GlpR and the PTS activation in haloarchaea has not yet been established.

Recently, the genome of H. mediterranei was completely sequenced by our laboratory (16). A genome-wide in silico analysis showed that the arrangement of fructose metabolism-related genes, including those of the PTS system in H. mediterranei, is identical to that of H. volcanii. However, in contrast to H. volcanii, H. mediterranei can synthesize biodegradable polymers such as polyhydroxyalkanoates (PHA) from many inexpensive carbon resources (17–19). Studying the mechanism of regulation of the PTS in H. mediterranei not only is useful to compare the different gene regulation strategies between bacteria and archaea but can also result in a deeper understanding of the carbon sensing and utilization by this specific haloarchaeal PHA producer. In the present study, a comprehensive investigation of the regulatory mechanism of the PTS was performed in H. mediterranei. We demonstrated that GlpR is an indispensable activator of the PTS gene cluster upon fructose induction via direct binding to the PTS promoter region (P_PTS). Interestingly, we also revealed an additional posttranscriptional mechanism which could increase the translation efficiency of PTS transcripts. Together, our results help elucidate the complex and delicate mechanisms of fructose PTS regulation in the domain of archaea.

MATERIALS AND METHODS

Strains and growth conditions.

The strains used in this study are listed in Table S1 in the supplemental material. Escherichia coli JM109 was used as the host for the cloning experiments (Novagen, Madison, WI, USA) and was grown in Luria-Bertani medium at 37°C (20). Unless otherwise noted, H. mediterranei DF50 (21) and the gene knockout mutants were cultivated at 37°C in nutrient-rich AS-168L medium (22), and H. mediterranei strains harboring expression plasmids were cultivated in AS-168SYL medium (AS-168L without yeast extract) (22). Chemically defined medium (CDM) [consisting of (per liter) 150 g NaCl, 20 g MgSO₄ · 7H₂O, 2 g KCl, 50 mg FeSO₄ · 7H₂O, 0.36 mg MnCl₂ · 4H₂O, 5 g NH₄Cl, and 15 g piperazine-N,N′-bis(2-ethanesulfonic acid) PIPES, pH 7.2] with different concentrations of fructose or glucose was used to verify the utilization of the carbon source by H. mediterranei mutant strains. When required, ampicillin, uracil, and 5-fluoroorotic acid (5-FOA) were added to the media at final concentrations of 100 mg/liter, 50 mg/liter, and 250 mg/liter, respectively.

Gene knockout and complementation.

In-frame deletion and complementation strains were generated according to previously published protocols (21, 23). All the primers used in this study are listed in Table S2 in the supplemental material, and the plasmids are listed in Table S1 in the supplemental material. The transformation of H. mediterranei was performed by the polyethylene glycol-mediated spheroplast transformation method (24). The plasmid sequences and mutant strains were verified by PCR and DNA sequencing.

RNA extraction, quantitative reverse transcription-PCR (qRT-PCR), and circularized RNA (CR)-RT-PCR.

H. mediterranei DF50 cells and the gene knockout mutants were cultured at 37°C in AS-168L medium. When the optical density at 600 nm (OD₆₀₀) reached 1.5, glucose or fructose was added to the medium to a final concentration of 50 mM, and the cells were incubated for 45 min. The sugar-induced cells (3 ml) were then immediately collected for RNA extraction using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. An equal volume of TBSL buffer (consisting of [per liter] 150 g NaCl, 20 g KCl, 5 g MgSO₄ · 7H₂O, and 100 mM Tris-HCl, pH 7.1) (22) was added to the cells in the control group. To remove DNA contamination, DNase I (Promega, Madison, WI, USA) digestion was performed on 1 μg of diluted RNA.

The specific primer pairs for the target DNA regions (see Table S2 in the supplemental material) and suitable concentrations of the cDNA templates or genomic DNA were used for quantitative PCR (qPCR). The amplification and detection of target regions were performed on a Rotor-Gene Q real-time cycler (Qiagen, Valencia, CA) under a standard three-step PCR procedure (including initial denaturation at 95°C for 10 min followed by 40 cycles of denaturation at 95°C for 30 s, annealing at 55°C for 30 s and synthesis at 72°C for 30 s; a melting curve was generated by linear heating from 70°C to 95°C over 25 min). For the synthesis of the cDNA, 200 ng of DNase I-treated total RNA was reverse transcribed via random hexamer primers by using the Moloney murine leukemia virus reverse transcriptase (M-MLV-RT) (Promega, Madison, WI, USA). DNase I-treated RNA (without reverse transcription) was used to check for genomic DNA contamination.

CR-RT-PCR (25, 26) was used to determine the 5′ untranslated region (UTR) of the PTS gene cluster. RNA circularization was carried out as described previously (25). Self-ligated RNA was reverse transcribed via random hexamers primers as described above. The cDNA was first amplified with a gene-specific primer pair, cRT1F and cRT1R, and a second PCR was performed to enhance the specificity by using an inner primer pair, cRT2F and cRT2R. The products of the second PCR were cloned into the TA vector pUCm-T (Sangon Biotech, Shanghai, China) according to standard procedures, and 15 clones from each RNA sample were analyzed by sequencing.

Constructs used for transformation of H. mediterranei.

For analysis of the promoter activity in vivo, a plasmid-based transcriptional reporter system using a soluble modified red-shifted green fluorescent protein (smRSGFP) (27) was constructed as previously described (28). All the plasmids used to transform H. mediterranei cells were derived from pWL502, and the details of their constructions are shown in the supplemental material. The plasmids pL117, pPR, and pPF were used to analyze the wild-type promoter activities of PTS (P_PTS), glpR (P_glpR) and fba (P_fba), respectively. To analyze the regulatory region of P_PTS, 5′ flanking deletion mutants of P_PTS (named pL93, pL77, pL56, pL24, and pGFP-0) and scanning site-directed mutants (from pM+1216 to pM7975) were transformed into H. mediterranei DF50 to detect the mutated promoter activity. The activities of GlpR and Myc-tagged GlpR were determined by introducing the plasmids pL117CR and pL117Rm, respectively, into the H. mediterranei DF50 ΔglpR strain. In the promoter or 5′ UTR swap experiment, pHSP, pUTR-M, pT8, and pT17 were generated by constructing fusions of the promoter and 5′ UTR regions between hsp5 (amplified from the plasmid pSCM307 [29]) and the PTS gene cluster.

Analysis of the smRSGFP fusion reporter system.

The fluorescence intensity of each smRSGFP fusion reporter plasmid-harboring strain was measured using a Synergy H4 hybrid microplate reader (BioTek Instruments Inc., Winooski, VT, USA). The excitation wavelength was set to 488 nm, and the emission wavelength was 509 nm (27). All strains were incubated at 37°C in AS-168SYL medium until the OD₆₀₀ reached 1.5; 90 μl of each culture was then transferred to black polystyrene 96-well plates (3916; Costar, Corning, NY, USA). In most cases, 10 μl of a fructose or glucose stock solution was added to the designated induction wells on the 96-well plates to a final concentration of 50 mM. For tests of induction with the metabolic intermediate fructose-1-phosphate (F-1-P), the final concentrations of F-1-P and fructose were reduced to 2 mM. In each assay, 10 μl of TBSL buffer was used as a negative control for basal fluorescence intensity. All of the plates were incubated at 37°C for 8 h before measurement.

Primer extension.

H. mediterranei strains harboring different plasmids were used in the primer extension assay with a specific primer based on the sequence of the GFP gene. Primer extension reactions were performed using 5 μg of total RNA and 3 pmol of the 5′-biotin-labeled primer gfpRbio (see Table S2 in the supplemental material) with the reverse transcription protocol described above. The extension products were analyzed on an 8% acrylamide sequencing gel. A chemiluminescent nucleic acid detection module kit (Pierce-Thermo Scientific, Rockford, IL) was used for biotin detection.

Immunoprecipitation.

The interaction between GlpR and the promoter region was analyzed by the chromatin immunoprecipitation (ChIP) assay. Myc-tagged GlpR was expressed using the pRm plasmid (see Table S1 in the supplemental material) in the H. mediterranei ΔglpR strain. Cells were harvested at the mid-logarithmic phase (OD₆₀₀ = 1.5) with or without fructose induction. The ChIP experiments were performed according to a previously described protocol (30, 31). The enrichment of genomic fragments was analyzed by qPCR with the input DNA samples as controls. The primers used are listed in Table S2 in the supplemental material. The PCR and thermocycling conditions were the same as described above for qRT-PCR. Each ChIP assay had five biological replicates. In each ChIP sample, the enrichment of GlpR-Myc interacting with each locus was calculated compared to the input sample using relative quantitation.

Microarray assay and deep sequencing.

The RNA samples used for qRT-PCR from DF50 cells with or without fructose induction were subjected to microarray analysis. Oligonucleotide microarrays were designed and manufactured by Capital Bio and Agilent Technologies based on the whole genomic sequence of H. mediterranei. The microarray assay was carried out as previously described (32). Each assay was repeated three times. The resulting data were analyzed by Significance Analysis of Microarrays (SAM) software version 2.23b (33). The same protocols used for the microarray assays were also performed on the total RNA samples from the ΔglpR mutant strain with or without fructose induction. The deep sequencing of the transcriptome of H. mediterranei was performed on HiSeq sequencing systems (Illumina HiSeq 2000) at the Beijing Institute of Genomics of the Chinese Academy of Sciences.

Prediction of RNA secondary structures.

The program Sfold (34, 35) was used for the prediction of putative secondary structures of RNA (http://sfold.wadsworth.org).

Microarray data accession number.

The microarray data have been deposited in the NCBI GEO library under accession number GSE41134.

RESULTS

GlpR is essential for the activation of the PTS promoter by fructose.

In H. mediterranei, the PTS genes (HFX_1559 to HFX_1563) were organized in an operon corresponding to a polycistronic transcript (Fig. 1A). The functional involvement of the PTS in fructose utilization was confirmed via genetic methods that were used in H. volcanii (21) (see Fig. S1 and S2 in the supplemental material). The DNA sequence of the intergenic region between fruK and ptsC is given in Fig. 1B. Deep sequencing of the total RNA of H. mediterranei DF50 indicated that there were two transcripts, with 17- and 8-nucleotide (nt) 5′ untranslated regions (UTRs), respectively. The start site (A) (+1) of the longer transcript (T₁₇) is 9 nt away from the start site (G) (+10) of the shorter transcript (T₈) (Fig. 1B and C), and these start sites were confirmed by CR-RT-PCR. However, only one typical promoter containing a putative TFIIB recognition element (BRE) (−36 GAAAGG −31) and a putative TATA box (−30 ATTTTT −25) was found (Fig. 1B).

FIG 1.

Map of the PTS gene cluster, the promoter sequence, and analysis of PTS transcripts. (A) Genetic organization of the H. mediterranei PTS gene cluster and neighboring genes. glpR (HFX_1565) encodes a protein homologous to a DeoR family transcriptional regulator, and fruK (HFX_1564) encodes the 1-phosphofructokinase. ptsC, ptsA, hpr, ptsI, and ptsB (HFX_1563 to HFX_1559) encode a complete PTS, and fba (HFX_1558) encodes a fructose-1,6-bisphosphate aldolase. The primers used for CR-RT-PCR are indicated with arrows. (B) Promoter sequence of the PTS gene cluster is shown. The stop codon TAA of the upstream gene fruK and the start codon ATG of ptsC are boxed. The start sites (indicated by arrows) of the two transcripts containing 5′ UTRs of different lengths were determined by CR-RT-PCR. The putative TATA box and BRE are indicated by single and double underlines, respectively. (C) Statistical results of mRNA deep sequencing. The counts (y axis) of corresponding nucleotides (x axis) of a 24-bp sequence of the P_PTS are presented. The positions with the most significant increases in sequencing counts are marked by arrows, which indicate the two start sites of PTS transcripts.

In silico analysis showed that the DeoR family transcriptional regulator, GlpR, is highly conserved in haloarchaea (>60% identity), but it is more distantly related to its homologs in other archaea and bacteria (<45% identity) in comparison. The protein sequence multiple-alignment analysis by BLASTP indicated that approximately 70 of the 255 amino acids in the N terminus of GlpR form a putative helix-turn-helix (HTH) motif, and the remaining amino acids near the C terminus form a DeoR-type regulator C-terminal sensor domain. To determine whether the transcription of the PTS gene cluster is regulated by GlpR, H. mediterranei DF50 and the ΔglpR strain were analyzed by microarray assays with or without fructose induction (GEO accession number GSE41134). The transcriptional fold changes of the PTS gene cluster and neighboring genes (HFX_1558 to -1565) are listed in Table 1. It was shown that the PTS gene cluster and the glpR-fruK operon, which were highly upregulated by fructose in H. mediterranei DF50, were not inducible in the ΔglpR strain. In addition, the transcription of fba was not significantly changed in either group (Table 1). These results strongly suggested that GlpR is an indispensable regulator in fructose-induced PTS activation in H. mediterranei.

TABLE 1.

Comparative analysis of the transcriptional levels of the most relevant genes (HFX_1558 to HFX_1565) in H. mediterranei DF50 and the ΔglpR mutant during fructose induction, using a microarray assay

Gene	Annotation	DF50		ΔglpR mutant
		Fold change (mean ± SD)	q value (%)	Fold change (mean ± SD)	q value (%)
HFX_1558	Fructose-1,6-bisphosphate aldolase	0.82 ± 0.14	4.23	0.86 ± 0.11	10.58
HFX_1559	PTS IIB component	29.74 ± 7.31	0	0.92 ± 0.02	12.77
HFX_1560	PTS enzyme I	34.97 ± 14.17	0	0.88 ± 0.03	7.85
HFX_1561	PTS protein HPr	31.3 ± 12.77	0	0.85 ± 0.17	13.33
HFX_1562	PTS IIA component	40.37 ± 16.45	0	0.94 ± 0	20.2
HFX_1563	PTS IIC component	31.97 ± 4.33	0	1.11 ± 0.08	17.89
HFX_1564	1-Phosphofructokinase	17.33 ± 5.74	0	0.91 ± 0.12	21.9
HFX_1565	GlpR family regulator	21.07 ± 2.42	0

To confirm the regulation of PTS by GlpR in vivo and to conveniently investigate the PTS promoter, the ΔglpR strain harboring different GFP-based reporter plasmids was investigated; the DF50 strain served as a positive control. It was shown that whether induced by fructose or not, the fluorescence intensity of the ΔglpR strain harboring pL117, which expressed smRSGFP under control of the wild-type PTS promoter (134 bp upstream of translation start codon of ptsC), was much closer to the basal fluorescence intensity of DF50 harboring pL117 (without fructose induction) (Fig. 2). The ΔglpR strain harboring pL117CR showed that the activation of the P_PTS by fructose was restored through the expression of GlpR (using its native promoter, P_glpR) in the ΔglpR strain (Fig. 2). Furthermore, the transcriptional activity of P_glpR was also analyzed via the reporter plasmid pPR, which expresses smRSGFP using P_glpR. The fluorescence intensity of the P_glpR fusion reporter system increased slightly (approximately 1.5-fold) when DF50 cells were induced by fructose but did not change when ΔglpR cells were tested (Fig. 2). In contrast, regardless of fructose induction, the fluorescence intensity was similar in DF50(pPF) and ΔglpR(pPF) transformants (both expressing smRSGFP with the P_fba promoter) (Fig. 2). These results confirmed that GlpR is essential for the fructose-induced transcriptional activation of the PTS and the glpR-fruK gene clusters, and it may act as a positive regulator for fructose-induced PTS expression. This finding is quite interesting, as GlpR has been previously identified as a global repressor that inhibits the activities of KDGK and PFK in H. volcanii when cells are cultured in a glycerol-based medium (13). The different functions of GlpR are likely attributable to the different carbon sources (fructose versus glycerol) being used and/or to different promoters.

FIG 2.

GFP expression profiles of the DF50 and ΔglpR strains harboring smRSGFP-based reporter plasmids with or without fructose induction. The plasmids pL117, pPR, and pPF were constructed to express smRSGFP with the promoters P_PTS, P_glpR, and P_fba, respectively. Another construct, pL117CR, was transformed into the ΔglpR strain to express GlpR (using its native promoter, P_glpR) and smRSGFP (using P_PTS). The fluorescence intensity was detected as described in the text. Fructose was added to the cell cultures to a final concentration of 50 mM (Fru+). TBSL buffer was used as a negative control (Fru−). At least three independent experiments were carried out, and each experiment consisted of three replicates.

GlpR binds directly to the P_PTS during fructose induction.

To determine whether the activation of gene expression by GlpR occurs via direct binding to the promoter sequence, a ChIP assay coupled with qPCR analysis was performed on the ΔglpR strain harboring pRm (a Myc-tagged GlpR expression plasmid) with or without fructose induction. The recombinant GlpR-Myc (expressed by pL117Rm) was revealed to be able to restore the activity of wild-type GlpR in the ΔglpR strain (data not shown). Three DNA loci (F_PTS, F_iic, and F_phaE) were investigated for their interaction with GlpR, with the F_16S locus (a fragment of the 16S rRNA gene) used as an internal control for data normalization. F_PTS (119 bp) represented a fragment of the P_PTS region (bp −92 to +27 upstream of the PTS gene cluster), and F_iic (125 bp) represented the intragenic region of ptsC located approximately 550 bp to 650 bp downstream of F_PTS. The F_phaE locus (189 bp) containing the promoter region of phaE (which encodes a subunit of polyhydroxyalkanoate synthase in H. mediterranei [19]) was tested as a negative control because the expression level of phaE did not change in the microarray experiment when the cells were treated with fructose (data not shown). After the induction by fructose, the F_PTS locus exhibited a 2-fold enrichment of binding to GlpR-Myc over the negative-control locus F_phaE, and the enrichments of the F_PTS and F_phaE loci were similar to each other in the absence of fructose induction (Fig. 3). As expected, the fold enrichment of the F_iic locus remained unchanged and was similar to that of the F_phaE locus, with or without fructose induction (Fig. 3). These results demonstrated that GlpR could directly bind to P_PTS when H. mediterranei cells were treated with fructose, but the interaction between GlpR and P_PTS was not apparent without fructose induction. The significantly increased binding between GlpR and P_PTS under fructose induction indicated again that GlpR is an activator of PTS transcription.

FIG 3.

ChIP-qPCR data suggest that GlpR binds to DNA directly. The relative enrichment ratio of a 119-bp region of the P_PTS (F_PTS) immunoprecipitated by GlpR-Myc compared to randomly sheared chromosomal DNA (input samples), using a 145-bp region in 16S rRNA (F_16S) as an internal control to normalize the data, is shown. Enrichments are also compared for a 125-bp coding region (F_iic) of the ptsC gene and a 189-bp promoter region (F_phaE) of phaE that was not regulated by fructose induction in the microarray data. Cell samples for the ChIP assay were harvested under growth conditions with (Fru) or without (CK) fructose induction.

Three regions within P_PTS account for fructose induction.

The above results demonstrated that GlpR acts as a positive regulator for the induction of the PTS gene cluster by fructose via direct binding to the P_PTS. To experimentally analyze the cis-acting elements of the P_PTS, 5′ flanking deletion and site-directed mutagenesis analysis of the P_PTS were carried out based on the smRSGFP fusion reporter system of plasmid pL117. The fluorescence intensities of H. mediterranei DF50 transformants harboring deletions or site-directed mutagenesis constructs were measured with or without fructose induction (Fig. 4).

FIG 4.

Deletion analysis and site-directed mutagenesis of the P_PTS region. (A) Schematic representations (not to scale) of constructs pL117, pL93, pL77, pL56, and pL24. The wild-type P_PTS and truncated 5′ flanking sequence promoter mutants (solid lines, bp −93 to +17 to −24 to +17) were fused with the smRSGFP reporter gene (gray arrow). (B) Site-directed mutagenesis from bp −79 to +16 of the P_PTS. The DNA sequence of wild-type P_PTS is shown at the top (pL117). The two transcripts of the reporter gene with 5′ UTRs starting from bp +1 and +10, which are the same as observed in the PTS transcripts, were identified by CR-RT-PCR. The mutated nucleotides of different mutants (M7975 to M+1216) are shown below the wild-type (WT) promoter sequence. The basal (noninduced) and fructose-induced transcriptional activities of these promoters as revealed by fluorescence intensity were detected using a microplate reader. The basal transcriptional activities are expressed as a percentage of the pL117 activity (set to 100%), and the fructose-induced fold changes were calculated using the fluorescence intensity. Mutants that cannot respond to fructose are indicated with an asterisk, and mutants with a fluorescence intensity that was hardly detectable are marked by a dash. The significant fold changes in strains M+38 and M+1216, marked with a rhombus, were caused by the large decreases in the basal fluorescence intensity. At least three independent experiments were carried out, and each experiment consisted of three replicates.

The wild-type PTS promoter in pL117 and the deletion mutations in pL93, pL77, and pL56 exhibited similar basal transcription activities when cells were grown in AS-168SYL medium (data not shown). However, in the presence of fructose, the pL56 mutant completely lost the ability to respond to the fructose induction. The pL93 and pL77 mutants showed a 2- to 3-fold increase in fluorescence intensity after the fructose induction. When the putative BRE and TATA box of P_PTS were deleted (pL24), the transcription activities under both the basal and fructose-inducing conditions were almost undetectable (Fig. 4A). These results revealed that the pL77 mutant still contains the main cis-acting elements that respond to fructose induction. Thus, the promoter region from bp −79 to +16 of pL117 was analyzed using site-directed scanning mutagenesis to pinpoint the essential regions that account for the fructose induction. The resulting constructs were named pM7975 to pM+1216, in which the numbers indicate the mutation region. (For example, pM7975 indicates mutagenesis from bp −79 to −75 relative to the transcription start site [TSS] of P_PTS, and pM+1216 indicates mutagenesis from bp +12 to +16. These plasmids were transformed into H. mediterranei DF50 to generate the reporter strains M7975 to M+1216 for the detection of fluorescence intensity.

It was observed that mutations in the putative BRE (bp −36 to −31) and TATA box (bp −30 to −25) regions (M3935, M3430, and M2925) and the −10 region (M1410) led to a complete loss of transcriptional activity. Only three mutants, M7975, M6864, and M2420, had the same response to fructose induction as the pL117 transformant, and a 1.5- to 2-fold induction was detected in mutants M6360, M5553, M4440, and M1915 (Fig. 4B). It is noteworthy that the basal fluorescence intensity of M1915 rose to a very high level (19 times that of DF50 harboring pL117). However, the fluorescence intensity of M1915 was still induced by fructose, and thus the region from bp −19 to −15 might not be directly involved in fructose activation. GlpR seemed unrelated to any inhibition at the region from bp −19 to −15, since the fluorescence intensity of the ΔglpR strain harboring pL117 (with or without fructose induction) was similar to the basal intensity of DF50 cells harboring pL117 (without fructose induction) (Fig. 2). These results showed that the knockout of GlpR cannot enhance the P_PTS activity to as high of a level as that detected in the P_PTS-mutated plasmid pM1915. We speculate that either the site from bp −19 to −15 is required for the binding of an unknown inhibitor or the mutation from bp −19 to −15 changes the promoter architecture, both of which may lead to a higher activity of the mutated promoter under basal conditions.

Notably, the fructose induction did not significantly change the transcriptional levels of the mutants M7469, M5957, M5755, and M5351 to M4745 (the fold change for each was no more than 1.1), indicating that the corresponding regions in these mutants are important for the fructose-induced upregulation of the PTS. These results revealed that promoter regions I (bp −74 bp to −69), II (bp −59 to −56), and III (bp −52 to −45) are particularly important for fructose induction in H. mediterranei, and mutations in these three regions made P_PTS lose its ability to respond to the fructose induction (Fig. 4B). This phenomenon, which was similar to that observed in the ΔglpR mutant strain harboring pL117 (Fig. 2), indicated that regions I, II, and III were essential for the cellular responses to fructose induction and were likely to be the GlpR binding sites under fructose induction. Interestingly, after analyzing the P_PTSs of all six haloarchaea that possess the PTS gene cluster, a conserved 8-bp motif, which overlapped with seven base pairs of region III, was identified (Fig. 5). In addition, a palindromic DNA sequence pattern belonging to regions I and II, named motif P (short for “palindromic”) in this study, was also detected upstream of the 8-bp motif (Fig. 5), implying that the regulatory mechanism of the PTS revealed in H. mediterranei may be shared by other haloarchaea.

FIG 5.

Multiple alignments of promoter sequences of the P_PTSs in the haloarchaea which contain at least one complete PTS. Bases marked with asterisks are the sequences of regions I, II, and III. A palindromic DNA sequence (underlined, motif P) and an 8-bp sequence (boxed) are indicated. These sequences were found to be conserved via the alignment of the P_PTS in all promoter regions of haloarchaeal PTS.

It is noteworthy that the basal fluorescence intensity of mutant M+911 was more than 2.4-fold higher than that of the pL117 transformant, and both the basal and induced fluorescence intensities in the mutants M+38 and M+1216 decreased to a very low level in these strains (Fig. 4B). These results suggest that the mutations in the 5′ UTR altered either the mRNA stability or the translation efficiency of the PTS gene transcripts.

Generation of a PTS transcript with a shorter 5′ UTR due to posttranscriptional processing.

The results of scanning mutagenesis showed that only one TATA box, BRE, and −10 element were identified in the PTS operon (Fig. 4B), which indicated that there is only one promoter for PTS transcription. However, two transcripts with 5′ UTRs of different lengths (17 nt and 8 nt) were observed among the PTS transcripts (Fig. 1B and C), and the 5′ UTR of the PTS gene cluster was found to be important for their expression (Fig. 4B). To investigate how the two transcripts containing 5′ UTRs of different lengths were produced and whether they were involved in PTS activation, the features and function of this 5′ UTR region were further analyzed.

First, RNA folding and the general features of the 17-nt 5′ UTR were predicted using Sfold software, and a stem-loop structure was indicated. The start site (G) (+10) of the 8-nt 5′ UTR was located at the loop region (Fig. 6A). DF50 strains harboring different plasmids were used to conveniently characterize the 5′ UTR at both the transcriptional and translational levels (Fig. 6B). In a primer extension assay, when the nucleotides in the region from bp +3 to +11 (M+911 and M+38) of the 5′ UTR were mutated, the small extension product was hardly detectable (Fig. 6B). Therefore, the production of the shorter GFP gene transcripts likely depends on the sequence or structure of the 5′ UTR (same as the 5′ UTR of PTS transcripts) of the mRNA.

FIG 6.

Mutagenesis analysis of the 5′ UTRs of PTS transcripts. (A) In silico-predicted secondary structure of the 17-nt 5′ UTR of mRNA, (ΔG°₃₇ = −2.50). The designed mutation regions are marked with lines. (B) Electrophoretic analysis of the primer extension products of the GFP gene in strains M+911, M+38, and M1+2 after induction by fructose. DF50 cells harboring pGFP-0 were used as a negative control (NC), and DF50 cells harboring pL117 were used as a positive control (PC). The 59-nt and 41-nt oligonucleotides were used as molecular markers. The nonspecific products appearing in the NC are indicated with a and b, and the two primer extension products, corresponding to the transcript with the 17-nt (T_17G) or 8-nt (T_8G) 5′ UTR, are indicated with arrows.

To test this hypothesis, four promoter- or 5′ UTR-swapped constructs were generated, as shown in Fig. 7A. The plasmids, pT8, pT17, and pHSP contained the hsp5 promoter from Halobacterium sp. strain NRC-1 (P_hsp5) (36) and different 5′ UTR regions either from PTS (pT8 and pT17) or hsp5 (pHSP), respectively, whereas pUTR-M (containing the PTS promoter and the 5′ UTR of hsp5) was constructed in a manner similar to that for pHSP by only replacing P_hsp5 with P_PTS. The results showed that the two constructs that contained the 17-nt 5′ UTR, pT17 and the positive-control pL117, could both produce two transcripts (corresponding to T₈ and T₁₇ of the PTS transcripts and named T_8G and T_17G for the GFP gene transcripts) with different 5′ UTRs (Fig. 7B), despite the fact that the transcription of these constructs was controlled by different promoters (P_hsp5 and P_PTS, respectively). In contrast, only one transcript was generated from pHSP or pT8, as expected (Fig. 7B), and the PTS promoter (P_PTS) combined with the hsp5 5′ UTR sequence (pUTR-M) was also unable to generate the shorter transcript. These results indicated that the production of the transcript with shorter 5′ UTR was related only to the sequence of the longer 5′ UTR, and not to the promoter, and therefore indicated a potential posttranscriptional processing of PTS transcripts.

FIG 7.

Primer extension assay for identifying the different transcripts generated by P_hsp5-directed or P_PTS-directed reporter genes. (A) Sequences of constructs pT8, pT17, pHSP, and pUTR-M. The TSS (G) of P_hsp5 or the TSS (A) of P_PTS in the four constructs is marked with asterisks, and the different-length 5′ UTR of each construct is underlined. The sequence of the extension primer is boxed, and the expected size of the extension product is indicated in parentheses. The difference between pT8, pT17, and pHSP is the downstream sequence of the TSS (G), which is the 8-nt (pT8) or 17-nt (pT17) 5′ UTR sequence of the PTS or the 5′ UTR sequence of the wild-type hsp5 (pHSP), respectively. The only difference between pHSP and pUTR-M is that the P_hsp5 promoter is replaced by P_PTS in pUTR-M. (B) The primer extension products of pT8, pT17, pHSP (T₅₆), and pUTR-M were analyzed by electrophoresis. The primer extension products with same size are indicated with arrows (T₅₆, T_17G, or T_8G).

Physiological significance of the generation of the shorter 5′ UTR.

To understand the physiological significance of the production of two transcripts containing 5′ UTRs of different lengths, the translation efficiencies of these two transcripts were investigated. The relative amounts of GFP gene transcripts (detected by qRT-PCR) and their translation activities (measured by fluorescence intensity) were determined using DF50 strains harboring pT8 or pT17 (Fig. 8A). There was no significant difference in the amount of GFP gene transcripts between DF50 strains harboring pT8 or pT17 as evaluated by the Student t test, but the GFP expression from the shorter transcript in the DF50 strain harboring pT8 was over 6-fold higher than the GFP expression from the mixture of two transcripts in the DF50 strain harboring pT17 (Fig. 8A). These results indicated that the translation efficiency of the shorter transcript was much higher than that of the longer transcript.

FIG 8.

Functional characterization of the in vivo generation of the shorter 5′ UTR in H. mediterranei using a reporter gene. (A) The relative level of transcription activity and the translation efficiency of GFP gene transcripts in DF50 strains harboring pT8 or pT17 were determined via qRT-PCR and a fluorescence reporter assay. The levels of GFP gene transcription and the fluorescence intensity in strain DF50 harboring pT17 were both assigned a value of 1. At least three independent experiments were carried out, and each experiment consisted of three replicates. The statistical significance of the difference between the DF50 strains harboring pT8 or pT17 was analyzed using the Student t test. (B) Electrophoretic analysis of the primer extension products of the GFP gene in DF50(pL117) and ΔglpR(pL117) strains with (+) or without (−) fructose induction. Major transcripts are indicated on the left as described for Fig. 6.

Further research was performed to find a correlation between the fructose induction and the production of the two transcripts. The primer extension products of the DF50 and ΔglpR strains harboring pL117 with or without fructose induction were analyzed. The results showed that the amounts of the two transcripts increased as a result of fructose induction in the DF50 strain, and the ratio of the shorter transcript to the longer transcript was also increased (Fig. 8B). To confirm this result quantitatively, CR-RT-PCR was performed on the DF50 and ΔglpR strains with or without fructose induction. The ratio of clone counts of the PTS transcript with the 8-nt 5′ UTR to that of the transcript with the 17-nt 5′ UTR was doubled (from 14% to 31%) when DF50 cells were induced by fructose (Table 2). The ratios of T₈ to T₁₇ counts in the ΔglpR strain with or without induction were similar to each other and much closer to the noninduced ratio in the DF50 strain (Table 2). These results indicate that the translation efficiency of the PTS genes would be enhanced when the cells were induced by fructose due to the increased proportion of the shorter transcript.

TABLE 2.

Statistical results for the shorter and longer PTS transcripts in strain DF50 and the ΔglpR mutant with or without fructose induction

Strain	Fructose induction	Counts		Ratio (8-nt transcript/17-nt transcript), %
		8-nt transcript	17-nt transcript
DF50	−	5	36	14
	+	8	26	31
ΔglpR mutant	−	2	14	14
	+	2	20	10

F-1-P may acts as a positive intracellular effector.

During PTS regulation, the derivatives of sugars always serve as positive or negative effectors to enhance or repress the activity of regulators. F-1-P was shown to act as an important intracellular effector for the transcriptional regulation of the PTS in many bacteria (2, 14, 15, 37). To investigate the function of F-1-P during PTS regulation, the smRSGFP fusion reporter assay was performed with the pL117 transformant of H. mediterranei. In the fruK knockout mutant containing pL117, whether induced by fructose or not, the fluorescence intensity increased to a very high level (more than 20 times higher than that in DF50) and the induced expression of PTS by fructose disappeared (Table 3). In H. volcanii, it has been revealed that fructose is transported through the PTS, which would generate F-1-P, after which it is further phosphorylated by 1-PFK (encoded by fruK) (12). The high level of GFP gene expression in H. mediterranei ΔfruK may be caused by the accumulation of F-1-P when 1-PFK is inactivated, which implies that F-1-P may enhance the expression of the PTS as an intracellular effector as observed in bacteria. This hypothesis is also supported by the fluorescence intensity of H. mediterranei DF50 cells harboring pL117 when F-1-P is added to the culture medium (see Fig. S3 in the supplemental material). Either fructose or F-1-P induction significantly increased the fluorescence intensity, whereas the other derivative of fructose, fructose-1,6-bisphosphate (F-1,6-2P), decreased the fluorescence intensity (see Fig. S3 in the supplemental material).

TABLE 3.

Expression of the smRSGFP fusion reporter gene in H. mediterranei strains with or without fructose induction^a

Expt no.	Relevant host genotype	Relevant plasmid	Fluorescence intensity (mean ± SD)		Fold change
			Noninduced	Fructose induced
1	DF50	pL117	896 ± 158	5,052 ± 753	5.64
2	ΔfruK	pL117	22,108 ± 570	19,652 ± 612	0.89

^aAt least three independent experiments were carried out, and each experiment consisted of three replicates.

DISCUSSION

During fructose induction, GlpR has been shown to be an indispensable activator for the upregulation of the fructose-specific PTS gene cluster (Table 1; Fig. 2) through direct binding to P_PTS in H. mediterranei (Fig. 3). Therefore, GlpR is essential for the cellular responses to fructose induction. Interestingly, in H. volcanii, GlpR has been shown to be a global regulator by repressing the transcription of the key enzymes, including KDGK and PFK, when using glycerol as the carbon source (13). It seemed that the function of GlpR was decided by the environmental carbon sources. Through the regulator GlpR, glycerol represses sugar metabolism and fructose activates PTS expression. Previous reports suggest that the DeoR-type proteins always contain several highly conserved regions, one of which is the second helix of the helix-turn-helix (HTH) DNA binding motif in the N terminus (38). The other conserved regions are involved in oligomerization or inducer binding (in many cases, the inducer is a phosphorylated sugar). As one of the DeoR-type proteins, it makes sense that GlpR could be activated through the interaction of its C-terminal sensor domain with the fructose effector F-1-P in H. mediterranei. The activated GlpR (or GlpR accompanied by unknown regulators) could then bind to the promoter region through the N-terminal DNA binding motif to increase the transcriptional activity of P_PTS. This mode of action of GlpR was supported by the study of another DeoR-type transcriptional regulator, SugR, in Corynebacterium glutamicum (15, 39–41). Although the molecular details of the activation by GlpR require further investigation, one hypothesis, based on other transcription activation models in archaea (42–44), is that the activated GlpR recruits general transcription factors (transcription factor Bs [TFBs] and TATA binding proteins [TBPs]) to bind to the TATA box or BRE to enhance transcription.

According to a previous report, the possible binding sites of the GlpR as a repressor are located at the BRE or downstream of the TATA box as determined by in silico searching for the inverted repeat sequence (13). However, this kind of binding site that was reported for H. volcanii GlpR was not observed in the PTS promoter region in H. mediterranei. Instead, an upstream sequence (including regions I, II, and III) of P_PTS was indicated as the possible binding site of GlpR by the mutation scanning experiment (Fig. 4B). This finding showed a positional similarity with the upstream activator sequence (UAS) (from bp −52 to −39) of the bop gene in Halobacterium sp. strain NRC-1 (44–46). The conserved motifs (motif P and the 8-bp motif) in promoter regions I, II, and III of the PTS implied that the mechanism of PTS regulation by GlpR is similar in haloarchaea (Fig. 5).

It is noteworthy that the shorter PTS transcript would be generated through posttranscriptional processing in H. mediterranei (Fig. 7). This type of expression seemed to be different from its bacterial counterpart. Multiple TSSs were reported in the PTS genes of E. coli, which resulted from multiple promoters upstream of the coding sequences of the PTS genes and were influenced by DNA supercoiling and the transcription factor cAMP receptor protein (47), but the translation efficiencies of different transcript patterns were not very clear. Further analysis of the 5′ UTR sequences of T_8G and T_17G indicated that the relatively more efficient translation of T_8G was probably due to the presence of the shorter 5′ UTR (Fig. 8A). This hypothesis that the length of the 5′ UTR affects the translation efficiency of mRNA has also been given for other haloarchaea. In Halobacterium salinarum, leaderless mRNAs showed a higher translation activity than mRNAs with the Shine-Dalgarno (SD) sequence (48). On the other hand, the predicted stem-loop structure in T₁₇ (Fig. 6A) might also inhibit the reorganization or binding of the ribosome and hence repress the translation. It can be speculated that when induced by fructose, the translation efficiency of the PTS mRNA could be enhanced by increasing the ratio of T₈ to T₁₇ (Table 2). The relatively higher proportion of T₁₇ under conditions without fructose implied a constitutively low-level expression of the PTS.

The fructose-specific PTS of haloarchaea was thought to be acquired from bacteria by horizontal gene transfer (HGT) during evolution (12, 49). As a “gift from the neighbors,” the haloarchaeal PTS was also capable of “borrowing” the regulatory mechanism from bacteria at the transcriptional level. Furthermore, to acclimate to nutrient fluctuations in a competitive extreme hypersaline environment, the haloarchaea evolved their own mechanisms to control the PTS at the posttranscriptional and translational levels. Based on the results in this study and previous reports, we propose a working model for PTS regulation in haloarchaea (Fig. 9). In this model, fructose is transported into the cell and phosphorylated to F-1-P via the PTS and is further catalyzed to F-1,6-2P by 1-PFK. GlpR (or a combination of GlpR and other, unknown regulators) upregulates the transcription of this PTS gene cluster after the induction by fructose via direct binding to the P_PTS, most probably at the three conserved regions. F-1-P may act as the intracellular inducer, while F-1,6-2P may act as the negative effector, to be involved in this transcriptional regulation of PTS gene expression. A posttranscriptional processing of the PTS transcripts at the 5′ UTR, which increases the translational efficiency, is also involved in the PTS activation upon fructose induction (Fig. 9).

FIG 9.

A working model of the regulation of PTS expression and fructose utilization in H. mediterranei. CM, cytoplasmic membrane; 1-PFK, 1-phosphofructokinase; F-1-P, fructose-1-phosphate; F-1,6-2P, fructose-1,6-bisphosphate; TIC, transcriptional initiation complex.

In conclusion, the activation at the both transcriptional and translational levels would make the haloarchaeal PTS more efficient in response to environmental fructose. Although the working model has explained the main mechanisms of PTS regulation in haloarchaea, further study is warranted to determine whether other transcriptional regulators are involved in the regulation of PTS expression and to elucidate the mechanism of posttranscriptional processing. Such studies would help toward a comprehensive understanding of PTS regulation in haloarchaea.