INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The initiation of transcription of eubacterial genes is catalyzed by RNA polymerase and requires specific proteins, known as  factors. The main  factor used for transcription of housekeeping genes is ⁷⁰ (17), also known as ^A or ⁴³ in gram-positive bacteria such as Bacillus subtilis. The ⁷⁰ factor is essential for the viability of all known eubacteria (12). In Escherichia coli and B. subtilis, the ⁷⁰ promoter is characterized by two nucleotide sequences, TTGACA and TATAAT, respectively, centered at positions 35 and 10 relative to the transcription start site (+1). In more than 90% of the E. coli promoters these 35 and 10 hexamers are separated by 17 ± 1 nucleotides.

Promoters in gram-positive bacteria have other conserved regions in addition to the 35 and 10 hexamers. These include an A-rich region positioned at 43 and a region at 16 with the consensus sequence 5'-TnTG-3' (9, 16). At least part of the 16 region is also weakly conserved in E. coli ⁷⁰ promoters (1, 38). In general, the strength of a promoter is determined by how close its sequence resembles the consensus sequence and by the spacing between the 10 and 35 motifs (6).

Campylobacter jejuni is a major cause of bacterial diarrhea worldwide; however, compared to other enteric pathogens only a few genes have been cloned, and little is known about gene regulation in this bacterium (19). It is generally accepted that genes from C. jejuni are often difficult to clone or analyze in E. coli because of their high A+T content (70%). This may result in expression from normally nonfunctional promoter-like sequences and lack of specific expression due to the absence of required accessory factors (19). Compared to the ⁷⁰ promoter consensus sequences of E. coli or B. subtilis, the few housekeeping genes from C. jejuni that have been cloned, including the ones from which the transcriptional start sites have been mapped, show a typical 10 region but no 35 region or one that is very weakly conserved. With the exception of the C. jejuni glyA promoter, none of these putative promoters have proven promoter activity in C. jejuni or in E. coli (3). On the other hand, typical E. coli promoters like the lacZ or ampC promoters are not transcribed in C. jejuni (37). These anecdotal observations raise the question of whether an unusual promoter structure might occur frequently in the C. jejuni chromosome.

To determine promoter sequences active in C. jejuni, a direct approach in which genomic DNA fragments of C. jejuni were cloned in a shuttle promoter vector was followed. Fragments with promoter activity in C. jejuni were identified by their ability to cause expression of a promoterless lacZ gene. The nucleotide sequences of promoter-containing genomic fragments and the transcriptional start points of the transcribed mRNAs were determined. A consensus promoter sequence was deduced from the sequences of the various active promoters.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, plasmids, media, and growth conditions. C. jejuni 129108 (7) was originally isolated from a patient with recurrent C. jejuni infections and was obtained from H. P. M. Endtz, University Hospital Rotterdam-Dijkzigt, Rotterdam, The Netherlands. C. jejuni was routinely cultured at 42°C on Skirrow agar medium (34) under microaerophilic conditions (5% O₂, 10% CO₂, and 85% N₂). E. coli DH5 was used as the host for all plasmid constructions and transformations (31). Plasmids pILL550 (23), pBR322 (36), and pCB267 (33) were used for the construction of a promoter probe vector. E. coli strains were cultured in Luria broth (LB) or on LB agar. Antibiotics ampicillin (100 µg/ml) and kanamycin (30 µg/ml) were added when required.

Construction of a shuttle promoter probe vector. E. coli promoter probe vector pCB267 contains a promoterless lacZ gene. The DNA containing this gene was removed with restriction endonucleases BamHI and PvuII and was ligated into plasmid pBR322 digested with BamHI and XmnI. The resulting plasmid, pMW9, was digested with PstI to give a 5.1-kb fragment containing the lacZ gene and the origin of replication for E. coli. Finally, this fragment was ligated to a 5.1-kb PstI fragment of plasmid pILL550 carrying the origin of replication for C. jejuni and a kanamycin resistance gene.

Construction of a C. jejuni promoter library. Chromosomal DNA of C. jejuni 129108, prepared as described previously (39), was digested with Sau3AI. After 3 h of digestion the endonuclease was heat inactivated. To remove the smallest chromosomal DNA fragments, the digestion mixture was precipitated with isopropanol. Plasmid pMW10 isolated from C. jejuni 129108 with the Qiagen plasmid kit (Qiagen Inc., Chatsworth, Calif.) was digested with BglII and dephosphorylated with alkaline phosphatase (Pharmacia, Uppsala, Sweden). After phenol-chloroform extraction and ethanol precipitation, 2 µg of digested pMW10 was ligated with 3 µg of the remaining chromosomal fragments with 10 U of T4 ligase at 16°C for 16 h. The promoter library was obtained after electroporation of the ligation mixture into C. jejuni 129108. Electroporation was performed with a Bio-Rad Gene Pulser (Biotechnologies and Experimental Research Inc., San Diego, Calif.) set at 12.5 kV/cm, 25 µF, and 600 . Bacteria were suspended in 200 µl of heart infusion broth (Difco) and plated onto heart infusion broth agar plates. After 3 h of regeneration at 37°C, they were harvested and plated onto Campylobacter-selective Skirrow agar plates supplemented with kanamycin.

Screening for promoter elements. C. jejuni transformants were picked and grown in 96-well plates filled with 100 µl of 0.4% thioglycolate agar. After growth of the bacteria for 16 h at 42°C the -galactosidase activity was demonstrated by adding 50 µl of 0.1% 5-bromo-4-chloro-3-indolyl--galactose (X-Gal) in phosphate-buffered saline (PBS).

DNA sequencing and analysis. Plasmid DNA of C. jejuni was purified by the alkaline lysis method (31), with the modification that C. jejuni transformants were harvested from Skirrow agar plates after 16 h of growth with 1 ml of PBS containing 100 mM EDTA. Plasmids isolated from C. jejuni were electroporated into E. coli DH5. Plasmids from E. coli were isolated with the Qiagen plasmid kit. The number of bacteria in each E. coli or C. jejuni culture and the amounts of plasmids isolated from these cultures were calculated by measuring the absorbance at 600 or 260 nm, respectively. The sequences of the cloned DNA were determined by the dideoxy chain termination method (32), with an Autoread sequencing kit using T7 DNA polymerase (Pharmacia). Either a Cy5-labelled universal primer (UP) or a kanamycin primer (Km) was used (Table 1) in an automated laser fluorescent DNA sequencer (Pharmacia). PC/Gene, version 6.70 (22), was used to analyze nucleotide and amino acid sequences. The program Multalin, version 4.0 (5), was used to align promoter sequences. The symbol comparison table, identity, and gap weights were set on 1. 

RNA isolation and primer extension reaction. From each C. jejuni transformant total RNA was isolated from bacteria which had been grown for 16 h on thioglycolate plates. RNA isolation and cDNA synthesis were performed as described previously (40), with the exception that 50 U of reverse transcriptase were used. The primer used to map the 5' ends of the mRNA was the UP or one of the internal primers 1B7, 1G9, 4C7, 4C11, 5G10, 12G7, and 23E5, which are complementary to the coding strand (Table 1). The cDNA product was analyzed by electrophoresis on a 6% polyacrylamide gel containing 8 M urea, and its sequence was compared to sequence ladders obtained with the same oligonucleotide primer used for the primer extension. The sequence reactions were performed according to the dideoxy chain termination method (32).

-galactosidase assay. -galactosidase activity in C. jejuni and E. coli was measured by the conversion of o-nitrophenyl--D-galactopyranoside in nitrophenol as described by Miller (25), with the modification that C. jejuni transformants were grown for 16 h on thioglycolate plates before they were harvested with medium A (31) and diluted until the absorbancy at 600 nm was 0.4. Assays of -galactosidase activity were carried out in triplicate.

Nucleotide sequence accession numbers. The complete nucleotide sequence of pMW10 has been submitted to the EMBL nucleotide sequence database under accession no. AJ001494. The nucleotide sequences of the cloned C. jejuni promoter regions are available from the EMBL database under the accession numbers listed in Table 2.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Construction of a shuttle promoter probe vector. Shuttle promoter probe vector pMW10, designed and constructed for this study (Fig. 1), replicates in both C. jejuni and E. coli. It contains a promoterless lacZ gene 30 bp downstream of a small multiple cloning site. Translational stop codons between the multiple cloning site and the lacZ gene are present in all three reading frames. The lacZ gene is also preceded by a ribosome binding site (RBS) sequence. Endogenous expression of the promoterless lacZ gene on pMW10 was not detectable in C. jejuni or E. coli. The -galactosidase encoded by lacZ can be estimated colorimetrically in liquid cultures or in situ in colonies of C. jejuni and E. coli. The C. jejuni origin of replication (ORI) of pMW10 was sequenced. Genes mob and repB appear to be essential for replication in C. jejuni (23). The sequence of the ORI is similar to the sequences found in two cryptic Campylobacter coli plasmids registered under accession no. X82079 and X82080 (34a).

View larger version (24K): [in this window] [in a new window]   FIG. 1.   Map of promoter shuttle vector pMW10. This plasmid contains a Campylobacter-derived kanamycin resistance gene (km), the genes mob and repB, which are necessary for replication in C. jejuni, and the ORI (ori) for E. coli. A small polylinker region is situated in front of the promoterless E. coli lacZ gene. Restriction sites and the direction of transcription of the genes are indicated. Campylobacter sequences are represented by the thick-lined part of the circle.

Construction and screening of a promoter library. C. jejuni 129108 (7) was chosen as the host for the promoter library. It is a human isolate that can be transformed by electroporation with plasmids isolated from E. coli. Electroporation of C. jejuni 129108 with pMW10 isolated from E. coli was, however, less efficient (600 colonies/µg) than electroporation with the same plasmid isolated from C. jejuni 129108 (5 × 10⁴ colonies/µg). Digestion of C. jejuni 129108 chromosomal DNA with Sau3A resulted in DNA fragments with an average length of 500 bp. These fragments were ligated into pMW10, and the ligation mixture was introduced into C. jejuni by electroporation, resulting in 14,000 kanamycin-resistant transformants. We picked 1,824 colonies and screened them for -galactosidase. Direct screening on medium containing X-Gal inhibited the growth of transformants. Therefore the X-Gal was added after 16 h of growth. To prevent contamination between colonies by swarming, this had to be done in 96-well plates. Among the 1,824 transformants 125 showed varying degrees of blue color, which indicates that they contain a DNA fragment with promoter activity in front of the lacZ gene.

Sequence analysis of promoter elements. The plasmids of 11 C. jejuni transformants containing relatively strong promoters (dark blue colonies) were isolated and transferred by electroporation into E. coli. The amounts of the different plasmids isolated from C. jejuni transformants or E. coli transformants, starting with approximately 10¹⁰ bacteria, were each 5 ± 2 µg. This indicates that the cloned DNA fragments in pMW10 did not alter the copy number of pMW10 in these organisms. The sequences of the inserts in these plasmids were determined. Each clone contained a different chromosomal DNA fragment with an open reading frame (ORF) preceded by a typical RBS (Table 2, Fig. 2). In five cases, homology of the translated products of these ORFs with protein sequences in the databases was found (Table 2). Clone 1B7 contains an insert for which the encoded protein showed homology with the hypothetical protein YabC of Mycoplasma pneumoniae. The translated product of clone 4C11 showed homology with the isocitrate dehydrogenase of other bacteria. Plasmid 5G10 contains an insert homologous to C. jejuni ORF L1, which has already been sequenced by Hani and Chan (14). The function of L1 is unknown. Strong homology with glu-tRNA genes of numerous other bacteria was found for the insert of plasmid 12G7. Plasmid 23E5 contains a part of the gene coding for S-adenosylmethionine synthetase.

View larger version (49K): [in this window] [in a new window]   FIG. 2.   Alignment of 21 experimentally determined C. jejuni promoter sequences. The underlined nucleotide downstream of the 10 region for each sequence is the experimentally determined transcriptional start site. The consensus sequence derived from this alignment is given at the bottom. It consists of nucleotides that are present in any given position in more than 51% of the sequences. Dots indicate gaps introduced to maximize the alignment. The distances between each RBS and the associated transcriptional start site and between the transcriptional start site and the start codon of the downstream-situated gene are indicated (Gap). The origins of the promoter sequences are indicated in the "Ref." column. "a" stands for this study; the other letters and associated reference numbers are as follows: b, 30; c, 21; d, 18; e, 26; f, 10; g, 24; h, 3; i, 4. Promoter sequence nucleotides that match those of the consensus sequence are capitalized and in boldface; those that do not are lowercase and in lightface.

Determination of relative promoter strength. To determine the strengths of the 11 promoters described above, the -galactosidase activity resulting from the fragments containing these promoters was measured both in C. jejuni and E. coli (Table 2). C. jejuni and E. coli strains harboring plasmid pMW10, which does not produce significant -galactosidase activity, were used as negative controls. The -galactosidase activities of the 11 C. jejuni promoters varied more in E. coli than in C. jejuni. Only 6 of the 11 C. jejuni promoters showed detectable -galactosidase activity in E. coli. Reintroduction of the other five plasmids in C. jejuni resulted in dark blue colonies, indicating that they had not mutated.

Identification of transcriptional start sites. To localize the 5' ends of the mRNAs of the 11 promoters, primer extension experiments were performed. cDNA was synthesized from total RNA isolated from the 11 C. jejuni transformants indicated in Table 2, with the M13 UP (Table 1). In four cases this resulted in a cDNA product of which the start could be determined with single-base-pair accuracy. For the other 7 transformants internal primers (Table 1) which are located approximately 25 bp downstream of the putative start codons were used. For each promoter clone, we obtained a single dominant cDNA product. The RNA 5' ends identified in this way are underlined in Fig. 2. The average distance between the transcription start site and the putative initiation codon is 30 nucleotides.

Search for a ⁷⁰-like promoter consensus sequence. The 11 promoter regions characterized in this study as well as 10 C. jejuni promoters mapped in other studies (3, 4, 10, 18, 21, 24, 26, 30) were used to derive a promoter consensus sequence (Fig. 2). The initial alignment of the DNA sequences from position 50 to +1 of the promoter region was made by the program Multalin (5). Next, the alignment was visually inspected and adapted with respect to the known transcriptional start points. The promoter consensus sequence that could be deduced from this alignment contains nucleotides that appear in more than 50% of the cases at any given position. It consist of three regions: a 10, a 16, and a 35 region. The 10 region is positioned 4 to 11 bp before the experimentally determined RNA start points (Fig. 2). The consensus sequence of this region, TATAATT, is very similar to the sequence of the 10 region of a typical ⁷⁰ promoter of E. coli. In front of each 10 region, separated by one nonconserved nucleotide, the sequence TTTTTTTG, known as the 16 region (38), was found. This region is weakly conserved in E. coli but is more common in gram-positive bacteria (1, 9, 38). The 35 consensus sequence TTTAAGTnTT of C. jejuni completely differs from those of E. coli and B. subtilis. Only the sequences of the 3D8, 4C11, and hup promoters show some similarity to the 35 portion of the ⁷⁰ promoter consensus sequence of E. coli. The spacing between the 35 and 10 regions varies from 15 to 19 bp and is thus similar to the spacing observed in B. subtilis (16) and E. coli (15).

Putative RBSs in C. jejuni mRNA. In all cases a putative RBS was found, 1 to 62 nucleotides beyond the transcriptional start site (Fig. 2). Alignment of the 21 RBSs revealed that AAGGA was the consensus sequence. This consensus sequence is short compared to the C. jejuni 16S rRNA sequence from which an RBS with the sequence AGGAGG was predicted (20). The putative RBS is followed by a possible initiation codon, either ATG, TTG, or GTG, located after a spacer of 2 to 10 bp.

Comparison with other ⁷⁰ bacterial consensus sequences. The C. jejuni ⁷⁰ promoter consensus sequence was compared with those of B. subtilis (16), Corynebacterium glutamicum (28), Streptomyces spp. (35), and E. coli (15) (Table 3). The major C. jejuni consensus sequence is highly conserved compared to ⁷⁰ promoter consensus sequences of the other bacteria. There is a tendency that the number of conserved nucleotides, especially in the 10 and 16 regions of the ⁷⁰ promoter consensus sequence, increases when the G+C content of the bacterium decreases (Table 3). The conserved 35 region of the C. jejuni ⁷⁰ promoter totally differs from the 35 regions of other bacteria with the exception of the G at position 34.

View this table: [in this window] [in a new window]   TABLE 3.   Comparison of putative 70 consensus sequencesa

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The lack of typical E. coli ⁷⁰ promoter consensus sequences (15) in front of C. jejuni genes led to the question of what kind(s) of sequences act as promoters in C. jejuni. To answer this question we constructed promoter probe vector pMW10. This vector turned out to be an effective tool to identify promoter elements and allows subsequent testing of these elements in both C. jejuni and E. coli. Earlier studies concluded that the lacZ gene is not expressed in C. jejuni (37). From the results described here we can conclude that the lacZ gene, but not the lac promoter, is functional. The constructed promoter library in pMW10 is the first described chromosomal library made in C. jejuni. Theoretically, it contains all C. jejuni promoters, except the promoters containing a Sau3A site.

Eleven strong C. jejuni promoter elements were characterized. They induce various levels of -galactosidase activity in C. jejuni. A reason for these differences could be the variation in distance between promoters and the lacZ gene. The secondary structure of mRNA may affect both the rate of initiation of translation and the half-life of the mRNA molecules. Thus, direct measurement of product activity may not reflect actual promoter strength.

Another reason for the different levels of -galactosidase activity may be the sequence identity of a given promoter with the predicted consensus promoter sequence, which we assumed roughly correlates with the -galactosidase activity. Therefore, the 11 characterized promoter sequences and 10 promoter sequences from other studies were used to identify a potential C. jejuni promoter consensus sequence. Most promoters aligned in this study are situated in front of C. jejuni housekeeping genes which are generally assumed to be under the control of ⁷⁰ promoters (17). No clear correlation between promoter sequence and/or distance between the promoter and the lacZ gene and -galactosidase activity in C. jejuni was found.

The C. jejuni promoter consensus sequence, which we derived in this study, differs from the ⁷⁰ promoter consensus sequences of E. coli and other bacteria, especially in the 16 and 35 regions. The 16 region, mainly found in gram-positive bacteria, is extended (9, 38). In B. subtilis, mutations in the 16 region may result in a 94-fold decrease in transcription; the G residue at position 15 is particularly important (1, 38). Functional E. coli promoters lacking a conserved 35 region but containing a TG dinucleotide at positions 14 and 15 have been described previously (2). The function of the 16 region may be to compensate for weak 35 and 10 hexamers (38). The C. jejuni 35 region is highly conserved among the 21 compiled promoters, but its sequence is totally different from the other bacterial ⁷⁰ promoter consensus sequences.

The lack of expression of 5 of the 11 promoter regions in E. coli also indicates that the C. jejuni promoter sequence is different from that of E. coli. The results indicate that the lack of expression of C. jejuni genes in E. coli observed in earlier studies is probably due to the level of transcription (19). Five C. jejuni promoter regions, however, were found to be stronger in E. coli than in C. jejuni. This could be due to utilization of some other cryptic promoter sequences and/or alternative transcription start sites. Another explanation for this might be the different codon usage in E. coli and C. jejuni. For example, the 96 leucine amino acids present in the LacZ protein are encoded in 56% of the cases by CUG. In 71 proteins of C. jejuni only 1.1% of the leucine amino acids are encoded by CUG (Codon Usage Database of Genbank). Similar percentages are seen for other codons.

Allele-specific suppression of ⁷⁰ promoter mutations by specific changes in the corresponding  factor in E. coli resulted in the identification of two conserved regions which recognize the 10 and 35 promoter elements (6, 11). We hypothesize that the poor transcription of biosynthetic and housekeeping gene promoters of C. jejuni in E. coli is due to differences in the ⁷⁰ amino acid sequences for these regions. This idea could be tested by the identification and characterization of a C. jejuni housekeeping ⁷⁰-factor gene. We have cloned this gene. Preliminary sequence data show that the Campylobacter ⁷⁰ is highly similar to the ⁷⁰ protein of Helicobacter pylori but at the same time is highly divergent compared to the products of other rpoD genes.

Previous studies have mentioned that the recognition specificity of the RNA polymerase complex decreases with a decreasing A+T content of the chromosomal DNA of a bacterium (27, 28). The large number of conserved nucleotides in the major C. jejuni consensus sequence is in agreement with the high A+T content in this organism. However, the suggestion made by Gruber and Bryant (12) that typical promoters in all eubacteria should contain sequence motifs and spacing similar to those of the consensus promoter sequence for the ⁷⁰ factor of E. coli is premature.

The major C. jejuni consensus sequence derived in this study can be used to identify additional promoters in the C. jejuni genome, once the genomic sequence becomes available. So far, we have analyzed the recA, peb1A, and katA genes of C. jejuni of which the transcription start points have not been mapped. They do contain promoter sequences similar to the promoter consensus sequence we found (8, 13, 29). Our study shows that it is necessary to determine promoter sequences of bacteria experimentally, since they may contain sequence motifs different from those present in E. coli or B. subtilis.