![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
Department of Biochemistry, University of Bayreuth, 95440 Bayreuth, Germany
Received 12 October 2006/ Accepted 7 December 2006
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The regulation of plasmid copy number has been studied in detail for different bacterial plasmid families. Orf56 shows sequence similarity to CopG, a protein involved in copy number control of the bacterial plasmid pLS1 from the pMV158 family. CopG has been shown to bind to a 13-bp inverted repeat in the common promoter region (–35 box) of the copG-repB operon and to repress its own synthesis and that of the replication protein RepB (15). Additionally, an antisense RNA is needed as a regulatory element in pLS1. RNAII acts as a translational repressor by binding to the ribosome binding site of the repB mRNA (10, 13, 14).
Because of its small size, its stable maintenance in Sulfolobus cells without integrating into the host genome, and the fact that it does not cause any growth retardation, the multicopy plasmid pRN1 is well suited as a backbone for the construction of Sulfolobus-Escherichia coli shuttle vectors. However, up to now, information on this plasmid has been limited to bioinformatic analysis and protein function analysis. There is currently no experimental evidence on the mode of replication of pRN1, on a putative origin, or on a minimal replicon. The analysis of the transcripts of this plasmid will help to understand pRN1 in more detail and will perhaps also be useful for finding suitable interruption sites for the insertion of an E. coli replicon for the construction of a Sulfolobus-E. coli shuttle vector.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
RNA preparation. RNAs were prepared by a Trizol method using RNAgents (Promega). RNA preparations were digested with DNase I (Roche) at room temperature for 30 min and subsequently column purified (RNeasy Mini kit; QIAGEN). This combination yielded the best reproducibility of parallel RNA preparations (coefficient of variation of RNA concentrations, 13% [n = 4]). A crucial problem in the preparation of RNAs from plasmid-containing Sulfolobus cells is the large amount of plasmid copurifying in the first step of purification (Trizol method). This amount is reduced by the DNase treatment, but there is still a very low level of residual DNA contamination that cannot be eliminated even by a second DNase digestion at 37°C. Since the second DNase digestion step increases the variability in parallel preparations and does not yield DNA-free RNA preparations, only one DNase digestion step was performed. The residual DNA does not interfere with quantitative reverse transcriptase PCR (qRT-PCR) measurements because the level of DNA molecules compared to cDNA molecules is lower by several orders of magnitude. Nevertheless, a specific PCR product was sometimes also detected in control reactions containing only RNA without the RT step, but at considerably higher threshold cycle (CT) values than in the corresponding samples.
Reverse transcription. Reverse transcription was performed using Transcriptor reverse transcriptase (Roche) at 55°C to reduce the influence of secondary structures. Initially, an RNase inhibitor (Roche) was added to the RT reactions, but it was subsequently left out because it showed an inhibitory effect on the qPCR when the RT product was used at a 25-fold dilution in the qPCR reaction mix. The inhibition could be circumvented by using a 250-fold dilution, which was commonly used for all sample quantifications.
The influence of the input amount of RNA on the transcribed cDNA amount was tested in the range of 0.5 µg to 5 µg of RNA and found to be linear in that range, allowing the use of different input amounts of RNA.
qRT-PCR. qRT-PCRs were carried out in 96-well plates (25-µl volume) with an ABI PRISM 7000 sequence detection system (Applied Biosystems). SYBR green real-time PCR master mix (Epicenter) was used according to the instructions of the manufacturer, using ROX dye for signal normalization. The master mix and primers were premixed and dispensed into a 96-well plate, and appropriate dilutions of standards, controls, and samples were added. Primers (Metabion, Martinsried, Germany) were used at a concentration of 500 nM. An example of qRT-PCR results is shown in Fig. 1.
|
Standards and calibration. Standards were prepared from plasmid pUCpRN1, the pRN1 part was cut out, and the fragment was purified by agarose gel electrophoresis and extracted using the Wizard SV gel and PCR cleanup system (Promega). The concentration of the fragment was determined photometrically, and a stock dilution containing 109 molecules per µl was prepared and stored at –20°C. Fresh dilutions were prepared for every measuring period at concentrations ranging from 108.5 to 103.5 molecules per µl, yielding final standard concentrations after dilution into the well plate of 107.5 to 102.5 molecules per µl. For other target genes, similarly purified fragments were used.
Sample analysis was always accompanied by measurements of standards for calibration purposes. The CT values for standards were plotted against the logarithms of the numbers of molecules per µl contained in the standards, and the calibration function was derived.
Accuracy, precision, and reproducibility. The accuracy of the developed method is difficult to assess with original samples, as the amount of a target RNA cannot be known. To quantify the recovery of a transcript after the RNA purification and reverse transcription steps, an in vitro-transcribed RNA was spiked in three parallels into three Sulfolobus cell pellets, and subsequently, the RNA was prepared, reverse transcribed, and quantified. To obtain the test RNA, a 300-bp fragment of the kanamycin resistance gene from E. coli was cloned into pGEM-T (Promega) and transcribed using T7 RNA polymerase (AmpliScribe high-yield transcription kit; Epicenter). The resulting transcript was checked for integrity on an agarose gel, and the concentration was determined photometrically. The recovery of the spiked transcript was determined to be 60%, and the precision for the three parallels was 6% (coefficient of variation).
The precision of the qPCR step alone, as judged from the three replicates that were measured for every standard concentration for every primer pair, was always in the range of 1%, with the maximum coefficient of variation being 3.5% (Table 1) .
|
To assess the reproducibility of the procedure, several RNA transcript levels (orf56, orf904, orf80, orf90a, orf72, orf90b, and orf90a [4]) were quantified using cell pellets from different fermentations on different days (sampled at ODs of 1.3 to 1.4) that were extracted on different days and quantified using different standard preparations. The reproducibility was determined to be 24% (difference of two determinations divided by the mean).
Plasmid copy number determinations. Plasmid copy number was determined by qPCR as described above (with primers RT3_orf56_for and -rev). Plasmids were prepared by alkaline lysis and diluted 105- to 106-fold for the measurement. Three independent replicates were sampled and analyzed for each OD. In addition, the qPCR results were found to be consistent with copy number determinations carried out by comparing band intensities of plasmid dilutions on agarose gels.
Adaptor RT-PCR. During reverse transcription, an adaptor primer consisting of 20 nt complementary to the 3' end of the target RNA and a 20-nt adaptor sequence was incorporated into the cDNA molecule. In the subsequent PCR step, only a primer matching the upstream region of the RNA and a primer consisting solely of the adaptor sequence were used, circumventing the false-positive results caused by traces of residual plasmid DNA in the RNA preparations. The adaptor primer (0.5 µM in the RT reaction) was removed after the RT step by purifying the cDNAs with the Wizard SV gel and PCR cleanup system (Promega). Control PCRs without the adaptor sequence primer and without RT were always included. Primer sequences are listed in Table 2, and additional primers used to narrow down the transcript ends can be found in Table S1 in the supplemental material.
|
Primer extension. A primer (0.1 pmol) 5' end labeled with T4 polynucleotide kinase (approximately 150 nCi) and 20 to 50 µg of RNA were used in a reverse transcription reaction according to the protocol of the manufacturer (Transcriptor reverse transcriptase; Roche), using a 10-µl reaction mix. The enzyme was heat inactivated at 85°C for 5 min, and 10 µl of denaturing loading buffer was added. The sequencing reactions run alongside the primer extension reactions were prepared using a T7 sequencing kit (GE Healthcare). The influence of secondary structures, such as stem-loops, on the primer extension reaction was tested using a perfect 8+3 stem-loop structure situated directly downstream of the stop codon of orf904. A primer extension reaction in that region did not show any sign of impairment by this stem-loop (results not shown); thus, secondary structures are unlikely to cause false termination bands. Another test was conducted to exclude the possibility that residual DNA could cause additional bands. A primer extension reaction with RNA without DNase I treatment showed exactly the same bands as those obtained with DNase I-digested RNA.
Plasmid construction.
The promoter region of orf56 and an upstream sequence (positions 1892 to 2162 in pRN1 [accession number NC_01771]) were PCR amplified using primers 5'-TCCTAGGCTAAGCCCGCCCTGTCTAAC and 5'-CCATGGGTGGATCAAAATTGTATCCGC, introducing an AvrII site with the forward primer and an NcoI site with the reverse primer (underlined). The same promoter region plus the orf56 coding sequence (positions 1892 to 2382 in pRN1) was PCR amplified using the same forward primer and 5'-CCATGGACACCCCCGTTCTTCTTG as the reverse primer. Fragments were cloned into pGEM-T (Promega) and checked by sequencing. The inserts were cut out using the introduced AvrII and NcoI sites and then ligated into the preassembling vector pSVA10 (S. V. Albers and M. Jonuscheit, unpublished data) cut with the same enzymes, thus replacing the tf55
promoter region of pSVA10 with the orf56 promoter region or the promoter region plus the orf56 coding sequence. From these preassembly constructs, the promoter regions, together with the lacS reporter gene sequence, were cut out using BlnI and EagI and ligated into pMJ05 (1), yielding pProm and pPromOrf56.
Transformation of Sulfolobus cells. Conditions for the transformation of cells by electroporation followed the protocol of Schleper et al. (40), as specified in the work of Jonuscheit et al. (21).
Southern blots. Single transformants of S. solfataricus PH1-16 containing the reporter gene constructs were tested for the presence and copy number of free and integrated forms of the shuttle vector by Southern blotting. Genomic DNA was prepared from 1 ml of a logarithmically growing culture, using a Chemagenic DNA Bacteria kit (Chemagen, Baesweiler, Germany) according to the instructions of the manufacturer. After PvuII digestion, the fragments were separated in a 0.8% agarose gel, transferred to a Hybond N membrane (Amersham) by capillary transfer, fixed by UV irradiation for 5 min on a UV transilluminator, and hybridized to two digoxigenin-labeled probes complementary to positions 2580 to 3432 and 7455 to 7893 of Sulfolobus shibatae virus 1 (SSV1) (accession number NC_001338.1), indicating either the integrated or the free form of the shuttle vector construct. Labeling and detection were done using a PCR DIG probe synthesis kit and a digoxigenin labeling and detection kit (Roche).
ß-Galactosidase assay. ß-Galactosidase activity was measured from crude extracts prepared by a freeze-thaw method (21). Cells from a logarithmically growing culture were resuspended in 50 mM sodium phosphate buffer, pH 7, and subjected to five freeze-thaw cycles (–196°C/+50°C). After centrifugation for 30 min at 10,000 rpm, the supernatant was stored at –20°C or assayed directly. All ß-galactosidase assays were conducted in triplicate in a 75°C bench-top shaker. The reaction mix consisted of 3 µl of crude extract (or water for blanks) and 90 µl of 50 mM sodium phosphate buffer, pH 7, and the reaction was started by the addition of 7 µl of 12-mg ml–1 o-nitrophenyl-ß-D-galactopyranoside solution. Incubation was continued for 5 min before the tubes were cooled rapidly on ice, and 100 µl of 1 M Na2CO3 solution was added to stop the reaction. The concentration of o-nitrophenol was subsequently determined in a 96-well plate in a plate reader at 410 nm, using a standard curve generated with o-nitrophenol. The protein concentrations in the crude extracts were determined by the method of Ehresmann et al. (16).
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Initially, RT-PCR was used to deduce information on the lengths of the transcripts but proved to be not suitable because of residual DNA contamination in RNA preparations. Thus, an adaptor RT-PCR approach (see Materials and Methods) was used to circumvent this problem. Using this approach, we could show that orf56 and orf904 are cotranscribed, that the sense transcript of orf80 has a 3'-untranslated region overlapping with the antisense transcript ctorf90a, and that ctorf904 overlaps with the putative Shine-Dalgarno (SD) sequence upstream of orf904 (Fig. 2 and Table 3; also see Fig. 4).
|
|
The transcript comprising orf56 and orf904 starts 9 nt upstream of the ATG start codon for Orf56 at the last guanosine nucleotide preceding the start codon (Fig. 3). We also searched for a potential transcript originating from a starting point downstream of orf56, comprising only orf904, but we found no indication of such an additional transcript by primer extension analysis.
|
The starting points of the antisense transcripts ctorf904, ctorf90a, and ctorf90b were also determined. Within the coding region of orf904 and directly following the end of the coding region, two more transcription initiation sites in the antisense direction were determined by primer extension, with start sites at positions 3325 and 5088 in pRN1, starting with a C and a G nucleotide, respectively. In adaptor RT-PCR experiments, it was shown that the resulting transcripts do not run through orf904 but terminate earlier. Upstream of the mapped transcription start site for ctorf90b, within a region of approximately 100 nt, four more initiation sites in the antisense direction were detected from primer extension experiments. Also, for the cotranscript and the transcript ctorf90a, more than one transcription initiation site could be detected in the primer extension analysis. The clearest band closest to the primer was marked as the transcription start site in these cases.
In the upstream regions of the transcript start sites, BRE and TATA box-like elements were identified (Fig. 3B). BRE and TATA boxes of the cotranscript, the orf80 transcript, and ctorf904 show the highest similarity to the consensus sequence derived from different archaeal promoters (4). The identified TATA box is centered around nucleotides –26 and –27 preceding the transcription start site, as predicted by bioinformatic analysis (43). The BRE/TATA box for ctorf90a is only weakly similar to the consensus sequence. Comparable levels of transcripts are obtained in this case by multiple initiation sites.
Transcript endpoints. Transcript ends were estimated from the adaptor RT-PCR by using different primers spaced approximately 60 nucleotides around the supposed ends. The transcript end was assumed to be situated between the last primer that yielded a PCR product and the first primer that did not yield a product. Figure 4 and Table 3 show all compiled information on starts and ends of the transcripts. We crosschecked the results of the adaptor PCR with the primer extension results and with additional qRT-PCR experiments performed, e.g., for the 3'-untranslated region of orf80 (results not shown), and found all results to be consistent.
The regions of the supposed transcript ends were searched for terminator sequences. There is only very limited information on transcription termination in Sulfolobus. A pyrimidine-rich sequence has been identified in SSV1 as an important feature preceding transcript ends (37). Within the last 60 nt of each transcript, U-rich stretches can be identified. We used MFold (48) to predict the secondary structure of each transcript end. Stem-loop structures were found for every transcript (7- to 9-nt stem, 3- to 8-nt loop) (Fig. 5) but showed only limited thermodynamic stability, especially at 75°C.
|
|
|
|
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The orf56/orf904 cotranscript and the orf80 transcript are leaderless transcripts starting only 9 and 10 nucleotides, respectively, upstream of the ATG start codon. Sulfolobus has been predicted to have many leaderless transcripts from genome-wide analyses of transcription signals (42, 43). Our findings are also consistent with the report of the orf57/orf80 transcript of pIT3 (35), which is entirely leaderless and also starts with a G nucleotide.
The cotranscript has no 3'-untranslated region, as it stops within the next 60 nt following the orf904 stop codon. The transcript of orf80, in contrast, has a long (approximately 200 nt) 3'-untranslated region whose transcript levels are comparable to those of the coding region. The 3'-untranslated region and most of the coding region are overlapped by the antisense transcript ctorf90a. This transcript might play a regulatory role. The region downstream of orf80 is not as well conserved as the coding and upstream regions of orf80 (23). Additional qRT-PCR experiments with the corresponding region of pRN2 showed that a transcript running through orf81 (equivalent to orf80 in pRN1) and a transcript in the antisense direction downstream of orf81 are present. For Sulfolobus, a survey of noncoding RNAs found transcripts targeting the 3'-untranslated regions of certain mRNAs (41). It is worth noting that the level of the orf80 transcript is remarkably stable. It is the only transcript that does not show any changes in transcript level except for declining when cells are dying. This was observed for three different fermentations sampled at different times and also for different locations within the transcript and might indicate that the number of orf80 transcripts is tightly regulated.
The transcript ctorf904, found in the antisense direction within orf56/orf904, spanning about the first one-third of the orf904 coding sequence and overlapping with the putative SD sequences of orf904, could also be involved in copy number control of pRN1. This RNA could act as a translational repressor similar to RNAII of pLS1. In contrast to the short, 50-nt RNAII, ctorf904 is much longer, but like other regulatory RNAs of bacterial plasmids, it is shorter lived than the mRNA that it overlaps (10).
The transcript found in the antisense direction in the region of orf90b does not allow any conclusions to be drawn concerning its putative function.
We checked for the presence of putative open reading frames within the region of the countertranscripts. For ctorf904, two open reading frames of 207 bp to 258 bp could be identified (positions 875 to 618 and 864 to 658 in pRN1), and for ctorf90a, two open reading frames of 162 bp to 168 bp were found (positions 2781 to 2614 and 2509 to 2348). None of the putative open reading frames showed similarities to those encoding known proteins. We concentrated our transcript analysis on the region of pRN1 that is covered by open reading frames. For that reason, it is possible that further transcripts exist in the region between orf904 and orf80.
The stem-loop structures identified in the region of the transcript ends exhibit some similarity to bacterial rho-independent transcription terminators that consist of a stem of eight or nine, mostly G-C, base pairs and a loop of four to eight bases (45). Reiter et al. (37) suggested that the pyrimidine-rich sequence stretches found in the mRNAs of SSV1 resemble eukaryotic termination signals and could not identify secondary structures near the ends of the mRNAs. In contrast to the viral termination signals, the mapped transcripts of pRN1 clearly show stem-loop structures followed by a U-rich sequence. However, this sequence does not closely resemble the consensus of Reiter et al. (UUUUUUU/CU). Termination has been shown to take place at RNA hairpin structures followed by a U-rich sequence in a Methanothermobacter methanoautotrophicus in vitro transcription system, yet with varying degrees of efficiency (38).
In contrast to the quite dynamic behavior of the pRN1 copy number during fermentation, the levels of transcripts did not show large changes over different growth phases, except for declining in stationary/death phase. One possible explanation for this observation could be that the main part of plasmid copy number control is exerted not on the transcriptional level but on the translational or even posttranslational level. It has been shown that some transcripts from Sulfolobus have long half-lives compared to those of bacterial transcripts (8), and it was pointed out that this renders control on the transcriptional level difficult. It was proposed that control on the translational level could play an important role in archaea. The half-lives of the pRN1 transcripts (especially the cotranscript) are in the upper range of already determined half-lives for Sulfolobus (2, 8), and copy number control might operate on the translational level, perhaps involving the discovered antisense transcripts. For the bacterial plasmid pLS1, the main regulatory element is the antisense transcript RNAII, which responds quickly to copy number fluctuations, whereas the role of the CopG protein is to keep the synthesis of the copG-repB mRNA within narrow limits (14).
These findings are in accordance with a study on the transcription of Sulfolobus rod-shaped virus (SIRV1 and SIRV2), where a remarkably constant transcription pattern was observed upon infection of Sulfolobus cells (24).
The reporter gene assay proved that Orf56 does indeed repress transcription from its own promoter. The protein Orf56 has been shown in vitro to bind as a tetramer to a 12-bp inverted repeat situated just upstream of the orf56 start codon. Its binding region does not overlap with BRE or TATA boxes that are found 20 to 33 nt upstream of the ATG, and thus it cannot interfere directly with binding of either transcription factor B (TFB) or TATA box-binding protein (TBP) (32). However, the Orf56 binding site overlaps the transcript start, and thus the contact of this region and the RNA polymerase might be impaired. This bacterial mechanism of RNA polymerase abrogation has been demonstrated for the archaeal species Archaeoglobus (MDR repressor) and Pyrococcus (LrpA and PhrA) (5, 12, 44). In Sulfolobus, transcriptional repression can also occur by blocking of the binding sites for TBP and TFB (6), and it has been suggested that derepression takes longer than in the case of RNA polymerase abrogation, where TBP and TFB can stay bound to the promoter region. To prevent plasmid loss in case of downward copy number fluctuations, it is crucial that the repression of the replication operon can be reversed quickly.
The 11-fold repression at the transcriptional level is comparable to the 10- to 20-fold repression observed in the presence of the transcriptional repressor CopR for plasmid pIP501 (9), which is also regulated by a repressor protein and an antisense RNA. Similar to the case for CopR of pIP501 and CopG of pLS1 (10), the transcription of the cotranscript is not totally blocked in the presence of Orf56.
The reporter gene assay using ß-galactosidase activity yielded about two times more repression than the direct measurement of transcript levels. This could point to another mechanism at the translational level being active in the reporter gene shuttle vector constructs. It either favors the expression of the LacS protein in the pProm (without Orf56) construct or reduces its expression in the pPromOrf56 (with Orf56) construct. A mechanism that would act like the former could be the enhanced translational efficiency of a leaderless transcript (as in pProm) compared to initiation at an SD sequence (as in pPromOrf56). This has been observed for Halobacterium salinarum in vivo (39) and has also been described for a Sulfolobus solfataricus in vitro translation system (11).
The pRN family plasmids contain all homologs of Orf56/CopG and the replication protein Orf904 in their conserved regions. In the plasmids pTAU4, pTIK4, pORA1, and pIT3 as well as in conjugative plasmids and viruses of Sulfolobus, CopG-like proteins have been annotated preceding an open reading frame predicted to code for a replication protein (17). In some cases (e.g., pORA1 and pSSVx), the CopG-like proteins do not directly precede the replication protein but are followed by another short open reading frame encoding a protein also showing some similarity to Cop proteins (18). Inverted repeats as putative binding sites upstream of the CopG-like proteins have been identified in pKEF9 and pARN4 but not in pORA1 and pTIK4 (18, 19). The transcription of two CopG-like proteins preceding a replication protein has been examined so far only for pIT3. Two distinct transcriptional units consisting of a transcript for the CopG-like proteins and one for the replication protein (35) have been found, unlike the case in pRN1. Therefore, care has to be taken when transferring the results obtained with pRN1 to other annotated CopG-like/replication protein couples from Sulfolobus plasmids.
The combined information obtained from the biochemical characterization of the proteins encoded by pRN1, the analysis of its transcription pattern, and the results of first genetic experiments has considerably improved our understanding of plasmid organization and copy number regulation. The mechanism of copy number regulation shows similarities to that of gram-positive bacterial rolling circle plasmids of the pMV158 family, whereas the replication protein encoded in the autoregulated operon is a specific archaeal protein. Unexpectedly, in a large part of pRN1, transcriptional activity can be detected. This is reminiscent of the finding that transcripts are produced by at least half of the typical eukaryotic genome (7). The transcript levels are low, allowing the plasmid to stay inconspicuous without being a burden for the cell. Whether all of the transcripts have a function remains to be determined. Improvement of genetic systems for Sulfolobus will allow us to address still unresolved issues, such as the mode of replication of pRN1 or further mechanisms of regulation.
|
ACKNOWLEDGMENTS |
|---|
This work was supported by the DFG (grant Li913/3 to G.L.).
|
FOOTNOTES |
|---|
Published ahead of print on 15 December 2006.
Supplemental material for this article may be found at http://jb.asm.org/.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Appl. Environ. Microbiol. | Infect. Immun. | Eukaryot. Cell |
|---|---|---|
| Mol. Cell. Biol. | J. Virol. | Microbiol. Mol. Biol. Rev. |
| ALL ASM JOURNALS |
G values for the stem-loop structures are shown as predicted by MFold (
