INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Its extraordinary tolerance to extremely high doses of ionizing radiation has made Deinococcus radiodurans the focus of growing scientific interest. This non-spore-forming bacterium is able to survive up to 4,000 times the lethal radiation dose for humans without mutation or loss of viability (2, 9). D. radiodurans is also of interest as a representative of a deeply branching family within the domain Bacteria (10). The sequence of the D. radiodurans R1 genome was recently published and shown to consist of two chromosomes, a megaplasmid, and one plasmid (17).

Despite the interest in D. radiodurans, little is known concerning basic gene expression and promoters. Earlier studies showed that Deinococcus promoter regions are poorly recognized in Escherichia coli, and E. coli promoters that were tested were not recognized in D. radiodurans (7, 14), suggesting that deinococcal promoters might be different from the classical E. coli ⁷⁰ type. However, no transcriptional analysis of deinococcal promoters has been carried out. Analysis of the recently published genome sequence revealed only three putative sigma factors, one classing with vegetative ⁷⁰ (rpoD) sequences, and two classing with extracytoplasmic alternative transcription factors (annotated as rpoE and DR0804 [17]). Surprisingly, orthologs of the nitrogen-starvation, general starvation, and heat shock sigma factors (rpoN, rpoS, and rpoH, respectively) were not found.

One reason for the lack of information on promoters in deinococci is the lack of convenient genetic tools for studying promoters. A promoter cloning vector has been described (7), but it involves an antibiotic resistance reporter and is a large plasmid with limited cloning sites. Therefore, we developed a suite of integrative promoter-screening vectors that allow the screening and assessment of promoter regions in D. radiodurans based on lacZ and xylE as reporters. These vectors were used to isolate and analyze promoter regions, and promoter regions were further defined by transcriptional analysis. Surprisingly, the 10 and 35 sequences of these promoters are similar to the E. coli ⁷⁰ sequence.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. The bacterial strains and plasmids used in this study are listed in Table 1.

Chemicals and enzymes. All chemicals used were of analytical grade and, unless indicated otherwise, were obtained from Baker Chemical Co. (Phillipsburg, N.J.) or Fisher Scientific (Fair Lawn, N.J.). 5-Bromo-4-chloro-3-indolyl--D-galactopyranoside (X-Gal) and O-nitrophenyl--D-galactopyranoside (ONPG) were from ISC Bioexpress (Kaysville, Utah) and Sigma Chemical Co. (St. Louis, Mo.), respectively. Enzymes for molecular biology were purchased from Roche Molecular Biochemicals (Indianapolis, Ind.) and New England Biolabs (Beverly, Mass.) and used according to the supplier. Taq and Platinum Taq DNA polymerases were obtained from Gibco-BRL (Gaithersburg, Md.).

Media and growth conditions. Luria-Bertani (LB) broth for growth of E. coli consisted of (per liter) 10 g of tryptone (Difco Laboratories, Detroit, Mich.), 5 g of yeast extract (Difco), and 10 g of NaCl (pH 7.4). TGY broth for D. radiodurans contained (per liter) 5 g of tryptone, 1 g of glucose, and 3 g of yeast extract (10). Solid media were prepared by addition of 1.5% agar (Difco) to either LB or TGY broth. Where necessary, media were supplemented with the appropriate antibiotics, all of which were obtained from Sigma. Ampicillin (Ap) was used at 50 µg/ml for E. coli. Tetracycline (Tc) was added to a final concentration of 2.5 µg/ml for D. radiodurans. Kanamycin (Km) was routinely used at 50 µg/ml for E. coli and 8 or 4 µg/ml for D. radiodurans grown on solid and liquid medium, respectively. Transformations of E. coli were performed either using commercially available cells (JM109 and TOP10 from Promega, Madison, Wis., and Invitrogen, Carlsbad, Calif., respectively) or by the CaCl₂ method (13). D. radiodurans cells were transformed as described previously (7a).

DNA manipulations. Miniscale plasmid DNA preparations of E. coli were obtained as described by Sambrook et al. (13). PCR products were purified using the Qiaquick PCR purification Kit (Qiagen Inc., Valencia, Calif.). Northern blot analyses were performed according to Sambrook et al. (13). PCR-generated probes were labeled for hybridization with [-³²P]-dCTP (800 Ci/mmol; NEN Life Science Products, Boston, Mass.) using the Random Primed DNA labeling kit (Roche). Primers for PCR amplification, sequencing, and transcription start site mapping purposes were of varying length and were obtained from Gibco-BRL (Frederick, Md.) (Table 2). Radioactive sequencing reactions for primer extension analyses (see below) were carried out using the T7 Sequenase kit (Amersham Pharmacia Biotech, Piscataway, N.J.). Non-radioactive nucleotide sequencing was performed at the University of Washington's Department of Biochemistry DNA Sequencing Facility, using an ABI Prism 377 sequencer (PE Biosystems).

Sequence comparisons and predictions. Computational analyses of DNA and predicted amino acid sequences were performed using the using the following internet-based programs. Similarity searches were carried out using the Blast algorithms available at http://www.ncbi.nlm.nih.gov/BLAST/. Amino acid sequences for these searches were retrieved from the Colibri (http://bioweb.pasteur.fr/GenoList/Colibri/) and SubtiList (http://bioweb.pasteur.fr/GenoList/SubtiList/) webservers, dedicated to the E. coli and Bacillus subtilis genome sequences, respectively. Multiple alignments were performed using ClustalW (http://www2.ebi.ac.uk/clustalw/). The presence of possible signal peptidase I cleavage sites was analyzed using the parameters at http://www.cbs.dtu.dk/services/SignalP/; analysis of primary protein structure was performed using the ExPASy ProtParam tool available at http://www.expasy.ch/cgi-bin/protparam. Preliminary sequence data for D. radiodurans were obtained from the Institute for Genomic Research website at http://www.tigr.org.

Plasmid constructions. The pAY/K and pROBe series of promoter probe vectors were constructed as follows. First, a PCR fragment carrying the putative D. radiodurans R1 -amylase (1,4--D-glucan glucanohydrolase) gene (amyE) was cloned in pCR2.1 (Invitrogen). The amyE gene was subsequently transferred to the EcoRI site of pUC19 and disrupted by insertion of the pUC4K Km marker in the HincII site, yielding pAY/K1. After partial PstI and T4 DNA polymerase treatment, pAY/K1.1 through 1.3 were obtained, each lacking one of the three PstI sites of pAY/K1. By fusing pAY/K1.2 and pAY/K1.3, plasmid pAY/K2 was constructed, carrying a unique PstI site between the amyE segments. This site was subsequently used for the insertion of PCR fragments carrying either the Pseudomonas putida xylE, E. coli lacZ, or Aequorea victoria gfp gene flanked by PstI and NsiI sites, yielding pROBe1, pROBe4, and pROBe5, respectively. All these vectors contain a unique BglII site that was used for the introduction of D. radiodurans R1-derived promoter fragments that were amplified by PCR and cloned into pCR2.1 or pCR2.1-TOPO. A second series of vectors were constructed by deletion of an NheI-XbaI fragment from pROBe1 carrying the amyE (pAMYE1), lexA (pLEXA1), and groESL (pGROES1) promoter fragments and subsequent cloning of a pMTL23-derived multiple cloning site in a unique XmaIII site located behind these promoter fragments. In a final step, the E. coli lacZ gene was inserted in the BglII and SpeI sites of these vectors, generating plasmids pAMYE4Z, etc.

Detection of -amylase activity. Km^r D. radiodurans R1 transformants were streaked on TY agar (TGY from which glucose was omitted), supplemented with 1% (wt/vol) soluble potato starch (Sigma, S-2004). After growth at 30°C for 2 days, plates were placed at room temperature for an additional 5 days. Haloes around -amylase-producing colonies were visualized using an iodine solution consisting of 0.6% (wt/vol) KI and 0.3% (wt/vol) I₂.

-Galactosidase assays. Expression of the lacZ reporter gene in E. coli and D. radiodurans colonies was detected using X-Gal (40 µg/ml). Quantitative analyses of lacZ expression were performed according to Miller et al. (8). Cell extracts of D. radiodurans were obtained by passing concentrated cell suspensions through a French press at 1,000 lb/in² using a J5-598A laboratory pressure cell press (Aminco, Silver Spring, Md.). To prevent degradation of the reporter protein in cell extracts, a protease inhibitor cocktail (Complete Mini; Roche) was used. Alternatively, -galactosidase activity was measured in toluene-permeabilized cells as follows. Cells were harvested by centrifuging 1-ml culture samples at 16,000 × g for 2 min. The pellets were subsequently resuspended in 500 µl of Z-buffer (8) supplemented with lysozyme (25 µg/ml) and DNase I (50 ng/ml). After incubation for 30 min at 37°C, 20 µl of toluene was added. The suspensions were incubated for another 60 min at 37°C, and aliquots (20 to 200 µl) were taken to measure -galactosidase activity.

Northern hybridizations. Total RNA was isolated from a maximum of 5 optical density at 600 nm (OD₆₀₀) units of exponentially growing cells using the RNA Perfect kit (Eppendorf-5 Prime Inc., Boulder, Colo.). The isolates were subsequently treated with RNase I-free DNase I (Gibco-BRL) to remove residual contaminating chromosomal DNA. RNA samples (±8 µg) and molecular size RNA markers (0.24 to 9.5 kb; Gibco-BRL) were electrophoresed on MOPS (morpholine propane sulfonic acid)-formaldehyde-agarose slab gels at 6 to 7 V/cm. Subsequently, RNA was either stained in ethidium bromide (1 µg/ml) for 20 min and destained overnight in diethyl pyrocarbonate-treated H₂O or transferred to positively charged Hybond-N⁺ membranes (Amersham Pharmacia Biotech) and probed with ³²P-labeled PCR fragments, according to Sambrook et al. (13).

Transcription start site mapping. Transcription start sites were mapped by means of primer extension according to the ThermoScript cDNA synthesis protocol, using 10 µg of total RNA. Primers were labeled with [³²P]-ATP (6,000 Ci/mmol; NEN) using T4 polynucleotide kinase (Roche). In each case, primers for the reverse transcription reactions were 18- and 30-mers and were located at different positions relative to the start codon.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction and characterization of an integration vector. In order to develop a system for analyzing promoters in D. radiodurans at the same copy number as the chromosome, a chromosomal integration vector was developed based on recombination replacement within a nonessential gene. A gene predicted to encode -amylase was chosen as the target insertion site, as similar systems have proven successful in other bacteria (4, 6). Insertions in this gene, encoding the 1,4--D-glucan glucanohydrolase enzyme, provide a screenable phenotype, the formation of turbid haloes around amylase-producing colonies on starch-containing agar plates. Analysis of the D. radiodurans R1 partial genome sequence indicated that a 1,452-bp open reading frame (ORF) (DR1472) located on chromosome I was a likely candidate for an -amylase ortholog, with identities to the E. coli malS and B. subtilis amyE genes. The predicted protein contains a possible signal peptidase cleavage site (¹⁷AQAAP²¹), suggesting that it may be translocated across the bacterial membrane. The corresponding ORF was amplified by PCR and cloned in pCR2.1. The pAY/K (general insertion vectors) and pROBe (insertion vectors containing reporter genes; Fig. 1) series of plasmids were subsequently constructed as indicated in Materials and Methods. Transformation of CaCl₂-competent D. radiodurans yielded Km colonies at low frequencies when the plasmids were propagated in E. coli JM109 (Table 3). Loss of function could be visualized by the absence of halos around colonies grown for several days on TY agar containing 1% (wt/vol) starch.

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Schematic representation of the integration vectors and their unique restriction sites. The checkered and solid gray bars represent the 5' and 3' portions of the D. radiodurans R1 amyE gene, respectively; the solid bar indicates the pUC19-derived vector moiety. Antibiotic resistance markers (Apr and Kmr) are represented as shaded arrows. Fragments to be fused to the reporter genes xylE and lacZ (open arrow) were inserted in the BglII site indicated in bold type. Also indicated are the EcoRI sites used in the generation of pAY/K1 (see text for details). The vectors that were used for the integration and analyses of the promoters described in the present studies were different in their orientation of the Px::lacZ fusion (see Materials and Methods).

Assessment of promoter activities from selected genes. Analysis of the preliminary genome sequence generated a number of potentially interesting genes for promoter analysis. Three genes were chosen at this stage, amyE (since it was being used as an insertion site) and two that might be under stress control, groES (DRO606), and lexA (DRA0344) (5, 16). The latter two were chosen as candidates for future development of stress-induced expression systems. In D. radiodurans (as in most other bacteria), groES is located immediately upstream of groEL (DRO607) (17), and both are involved in heat shock response in other bacteria (5). Using the vectors described above, transcriptional fusions between putative promoter fragments and reporter genes present on pAY/K2 plasmids were constructed to assess the activity of these promoters. The fragment containing the groES regulatory region was chosen so as to include a possible ³²-like promoter sequence (15) located 157 bp upstream of the groES translational start site. For lexA and amyE, each promoter fragment tested included approximately 200 bp upstream of the predicted start codon. Two of the reporter genes tested (xylE and lacZ) were reliable for plate screening, but only lacZ gave detectable activity in in vitro assays. Therefore, lacZ was used as a reporter in experiments involving activity measurements.

View larger version (19K): [in this window] [in a new window]   FIG. 2.   Growth (solid lines) and -galactosidase expression (dotted lines; nanomoles per minutes per OD600) of strains RM15 (amyE PlexA::lacZ), RM16 (amyE PgroESL::lacZ), and RM18 (amyE::lacZ). The time point at which cultures were shifted to elevated temperatures is indicated by the arrow.

Random screening for promoter fragments. In a separate series of experiments, a random Sau3A genomic library of D. radiodurans R1 was constructed in pROBe1 in order to isolate promoter-containing fragments from the genome. After establishment of recombinants in E. coli, clones were pooled, transferred to D. radiodurans R1, and screened for catechol 2,3-dioxygenase activity on plates. Pools generating transformants showing activity were separated into single clones which were tested again for activity in D. radiodurans R1. Several promoter-containing fragments were identified by this approach. A few showing the strongest activity were sequenced, and among these was a fragment derived from a gene that shows considerable similarity with the malate synthase A gene (aceB) of E. coli. By analogy, we designated the corresponding ORF aceR. The 5' terminus of this gene is identical to an ORF (DRA0277) described by White et al. (17). This putative promoter fragment was included in further studies.

Northern blots and transcriptional start site mapping. In order to further characterize expression of these genes, Northern blots were carried out. Total RNA was isolated from cells at different points in the growth cycle (Fig. 3A). Subsequent Northern hybridization experiments identified transcripts for amyE and groESL (1.45 and 2.02 kb, respectively) with RNA from early-exponential-phase cells, but we were unable to detect a lexA transcript in any of the RNA preparations (Fig. 3B). For amyE and groESL, the mRNAs detected were the correct size for single-gene and two-gene transcripts, respectively.

View larger version (78K): [in this window] [in a new window]   FIG. 3.   RNA analyses in wild-type cells of D. radiodurans R1. (A) Formaldehyde gel electrophoresis of total RNA isolated from stationary-phase (lane 1), early-exponential-phase (lane 2), and mid-exponential-phase (lane 3) cells. As a marker, 4 µg of an RNA ladder was applied (lane M). (B) Northern hybridization using probes for amyE, lexA, and groESL. The expected transcript sizes are indicated at the right-hand side of the panel (in kilobases).

View larger version (96K): [in this window] [in a new window]   FIG. 4.   Transcription start site mapping of the groESL promoter by primer extension. The experiments were performed in either the presence (+) or absence () of reverse transcriptase, and the two panels show the results with two different primers (PE and PE30). The flanking sequences represent the upper (lightface) and lower (boldface) this is the strand shown) strands. The nucleotide used as the transcription initiation site is marked with an asterisk. An additional minor cDNA product () was obtained in both reactions; these could reflect an alternative transcription initiation site.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Despite the fact that D. radiodurans R1 has been the focus of increasing scientific interest over the past decade, only a limited number of molecular tools for genetic engineering of this radioresistant bacterium are currently available and very little is known about promoters and expression. The present paper describes the development of a series of double-crossover integrative vectors and their use in the cloning and initial characterization of promoters isolated from the D. radiodurans R1 genome and from a cryptic D. radiodurans SARK plasmid.

Although these integrative vectors were used in this study for promoter cloning, they are also useful for general cloning and expression in this strain. These vectors, in combination with the improved transformation protocols described here, will significantly enhance the ability to carry out genetic manipulations in Deinococcus strains.

Of the three potential reporter genes tested, only lacZ was successful for screening of colonies as well as for quantitative assays in cell lysates and in toluene-permeabilized cells. This system was used to identify three promoters with different strengths, one low level (lexA), one moderate level (amyE), and one strong (groESL). It was surprising to find a lexA ortholog in the D. radiodurans R1 genome sequence, because this strain was reported not to have an SOS-like error-prone response to DNA damage (10). Low-level reporter activity was obtained with the fragment tested, but the role of lexA in D. radiodurans R1 remains unclear.

Transcriptional start sites were mapped for five genes, including groESL. Alignment of the 10 and 35 regions of these sequences showed that two strong promoters (groESL and rpoBC) had 10 and 35 regions that were highly similar to the corresponding regions of the E. coli ⁷⁰ consensus promoter. This result was surprising because previous reports had suggested that Deinococcus promoters should be significantly different from E. coli promoters (7, 14). The other three regions upstream of mapped transcriptional start sites as well as an additional minor start site for groESL were more divergent, but the 10 and 35 regions still showed some similarity to the E. coli consensus sequence.

The two start sites mapped for groESL were about 60 bp apart. The minor (more upstream) start site overlapped a sequence with good similarity to the E. coli heat shock sigma factor (rpoH) recognition consensus sequence, which is involved in regulation of heat shock genes such as groESL in many bacteria (5). However, the 10 and 35 sequences upstream of this minor start site did not coincide with the same regions in the rpoH recognition sequence. The possibility that the transcription start site might be different under heat shock conditions was addressed. Although the cloned groESL promoter region was found to direct higher expression of the reporter gene under heat shock conditions, suggesting that D. radiodurans mounts a heat shock response, the transcriptional start sites were the same as in cells grown at normal temperatures. The basis for heat shock regulation in D. radiodurans is unknown, and further studies will be required to address this.

The present studies show that it is possible to efficiently insert heterologous genes into the genome of D. radiodurans using the vectors described in this paper. In addition, the constructs containing the various promoters described here can be used for insertional expression vectors with different levels of expression. Such new genetic tools will greatly enhance the ability to carry out genetic manipulations of D. radiodurans for a variety of basic and applied research applications.