|Home | About | Journals | Submit | Contact Us | Français|
Many human colonic Bacteroides spp. harbor a conjugative transposon, CTnDOT, which carries two antibiotic resistance genes, tetQ and ermF. A distinctive feature of CTnDOT is that its excision and transfer are stimulated by tetracycline. Regulation of the genes responsible for excision has been described previously. We provide here the first characterization of the regulation of CTnDOT transfer (tra) genes. Reverse transcription-PCR analysis of the region containing the tra genes showed that these genes are regulated at the transcriptional level. Surprisingly, increased production of tra gene mRNA in tetracycline-stimulated cells was mediated by the proteins encoded by the excision genes. Previous studies have shown that expression of the excision gene operon is controlled by the regulatory protein RteC. Accordingly, it was possible that RteC was also regulating tra gene expression and that the excision proteins were only accessory proteins. However, placing the excision gene operon under the control of a heterologous promoter showed that the excision proteins alone could activate tra gene expression and that RteC was not directly involved. We also found a second level of tra gene control. The transfer of CTnDOT was inhibited by a DNA segment that included only a portion of the 3′ end of one of the excision genes (exc). This segment contained a small open reading frame, rteR. By replacing the codons encoding the first two amino acids of the putative protein product of this open reading frame with stop codons, we showed that the rteR gene might encode a small regulatory RNA. RteR acted in trans to reduce the number of tra transcripts in a way that was independent of the excision proteins. The repressive effect of RteR was not the result of decreased stability of the tra mRNA. Instead, RteR appears to be modulating the level of tra gene expression in some more direct fashion. The complex regulatory system that controls and links the expression of CTnDOT excision and transfer genes may be designed to ensure stable maintenance of CTnDOT in nature by reducing the fitness toll it takes on the cell that harbors it.
An important contribution to the spread of antibiotic resistance genes between human colonic Bacteroides spp. is made by self-transmissible elements called conjugative transposons (CTns). CTns, also called integrated conjugative elements, are normally integrated into the bacterial chromosome (14). To transfer, they are first excised to form a double-stranded circular intermediate, which transfers to a recipient cell by conjugation and then integrates into the recipient chromosome (15). A family of CTns that is widespread in Bacteroides spp. is represented by CTnDOT. CTnDOT is a 65-kbp CTn that carries two antibiotic resistance genes, tetQ and ermF (18). In addition to mobilizing itself, CTnDOT can mobilize coresident plasmids and smaller integrated elements called mobilizable transposons. Some of these mobilizable elements have been shown to carry antibiotic resistance genes. The transfer genes on CTnDOT are central to all these activities, but little is known about their organization and regulation.
Both excision and transfer of CTnDOT are stimulated by tetracycline. The genes involved in regulation, excision, and transfer are depicted in Fig. Fig.1.1. The regulatory cascade that controls excision and transfer starts with a three-gene operon, tetQ-rteA-rteB. The tetQ gene encodes a ribosome protection type tetracycline resistance protein. RteA and RteB constitute a two-component regulatory system in which RteA is the sensor and RteB is the response regulator. The production of TetQ, RteA, and RteB is controlled at the translational level by translational attenuation (20, 26). RteB activates the expression of rteC at the transcriptional level, and RteC protein in turn activates the expression of genes involved in excision (13). Three of the genes required for excision are part of the operon orf2c-orf2d-orf3-exc (22). The orf3 gene is not essential for excision, but the other three genes are (4). The excision proteins Orf2c, Orf2d, and Exc participate with the integrase (IntDOT) to carry out the excision reaction, which produces the circular transfer intermediate (4).
An unusual feature of CTnDOT excision and transfer is that the excision proteins not only participate in the excision reaction but also appear to be involved in the regulation of transfer. Preliminary evidence for a role of excision proteins in the regulation of transfer came from a previous study of an 18-kbp transfer region of CTnDOT that contains essential mobilization (mob) genes and other transfer (tra) genes (Fig. (Fig.1).1). This 18-kb segment, when cloned into a nontransmissible plasmid to form pLYL72, made the plasmid self-transmissible. pLYL72 contained not only the tra genes but also the transfer origin (oriT) and three mobilization (mob) genes. In the absence of other regions of CTnDOT, pLYL72 transferred constitutively (yielding 10−5 to 10−6 transconjugants per recipient), regardless of whether or not tetracycline was present in the medium (11). If, however, the regulatory region (tetQ-rteA-rteB and rteC) and the excision operon (orf2c-orf2d-exc) were provided in trans, the transfer of pLYL72 was regulated (27). That is, in the absence of tetracycline, the transfer frequency was reduced to a background level (<10−8 transconjugants per recipient) whereas, in the presence of tetracycline, the transfer frequency rose to 10−3 to 10−4 transconjugants per recipient.
Although it was clear that orf2c, orf2d, and exc were all required to increase the transfer frequency of pLYL72 from 10−5 to 10−6 transconjugants per recipient to 10−3 to 10−4 transconjugants per recipient, it was not clear whether the proteins encoded in this region were directly responsible for the regulation of transfer gene expression or whether they participated indirectly to facilitate the action of a known regulatory protein, RteC. The excision region also had the inhibitory effect of reducing the transfer frequency of pLYL72, but the entire excision region was not required for the inhibitory effect. A 500-bp portion of the 3′ end of the exc gene was sufficient to cause the transfer frequency of pLYL72 to drop from 10−5 to 10−6 transconjugants per recipient to fewer than 10−8 transconjugants per recipient (27). In this 500-bp region, there was a small open reading frame (ORF), which was tentatively designated rteR. This ORF could have encoded a 66-amino-acid protein.
We report here that expression of the transfer genes (traA to traQ) is activated directly by the excision proteins, independently of RteC. We show further that the rteR region encodes a small RNA (RteR), which reduces transcription of the transfer genes when the excision proteins are not present.
Bacterial strains and plasmids used in this study are listed in Table Table1.1. Bacteroides thetaiotaomicron 5482 strain BT4007 contains a copy of CTnDOT in the chromosome. B. thetaiotaomicron strain BT4004 contains a copy of CTnERL in the chromosome. Bacteroides strain BT4001ΩQABC contains an integrated copy of the central regulatory region comprising tetQ-rteA-rteB and rteC. BT4001ΩQAB contains an integrated copy of the tetQ-rteA-rteB operon only; rteC is not present in this strain. BT4001ΩQABC and BT4001ΩQAB contain no other parts of CTnDOT. Bacteroides strains were grown initially in chopped meat (Remel, Lenexa, KS). Strains were transferred into tryptone-yeast extract-glucose medium (10) and grown in the presence of tetracycline (1 μg/ml; induced) or in the absence of tetracycline (uninduced). The antibiotic concentrations used were as follows: ampicillin, 50 μg/ml; cefoxitin, 20 μg/ml; chloramphenicol, 15 μg/ml; erythromycin, 10 μg/ml; and gentamicin, 200 μg/ml.
Transfer of the translational and transcriptional vectors into Bacteroides strains was done via triparental mating (26). The donor that carried the translational or transcriptional vectors was Escherichia coli DH5α/MCR, and the donor that contained the IncPα RP1 plasmid was E. coli HB101(RP1).
A translational fusion of the traA gene and the E. coli β-glucuronidase (GUS) reporter gene (uidA) was constructed. The upstream region of traA including the putative promoter region was cloned in frame with the ATG start codon of the uidA gene of pMJF2. The fusion PtraA-uidA was then cut out of pMJF2 by PvuII and cloned into the NruI site of pC-COW (Table (Table1)1) to form pGW40.5 (G. Whittle, unpublished data). This traA::uidA construct was then transferred into BT4001ΩQABC or BT4001ΩQAB via triparental mating, as described previously (26).
To make the orf2c operon independent of rteA, rteB, and rteC, a transcriptional fusion of the excision operon (orf2c) to the heterologous maltose-regulated promoter PsusA was made. The plasmid pGW53, which contains the native excision operon and promoter, was digested with BsphI and blunted. Next, the vector was digested with SstI to excise the native excision operon without its promoter. The plasmid pNLY1::PsusA, which contained the maltose-inducible promoter PsusA (5), was digested with SmaI and SstI. The excision operon, including the ribosomal binding site region, was then cloned into pNLY1::PsusA by using SmaI and SstI to create a transcriptional fusion of the excision operon that is independent of rteC and inducible by maltose. The plasmid was then digested with PvuII to excise the PsusA-orf2c operon region and cloned into the Bacteroides shuttle vector pAFD1 (Emr) to create the plasmid pGRW3, which is compatible with pGW40.5 (Cmr). A plasmid containing the functional excision region which is regulated by RteC (e.g., pKSO7) or the maltose-regulated pGRW3 was then transferred into BT4001ΩQABC or BT4001ΩQAB containing pGW40.5. Transconjugants were selected on gentamicin-chloramphenicol-cefoxitin for pKSO7 and gentamicin-chloramphenicol-erythromycin for pGRW3. Strains containing pKSO7 were grown with and without tetracycline. Strains containing pGRW3 were grown with maltose or glucose. Cells were then harvested by centrifugation and disrupted by sonication. GUS activity was evaluated as described previously (7).
There was a consensus Bacteroides promoter (2) 290 bp upstream of the first possible start codon for TraA. To ascertain that this was the location of the traA promoter, a nucleotide change (from TTTG to GTTG, where underlining indicates the change) was made. Site-directed mutagenesis in the putative promoter of the traA-uidA reporter construct pGW40.5 was performed using the QuikChange II site-directed mutagenesis kit (Stratagene). The effect of the mutation was measured by assessing GUS activity (7), and the GUS results were compared to those for the wild-type traA-uidA fusion.
To create a polar mutation in the tra operon, a PCR product that overlaps the 3′ end of traA (beginning 45 bp from the 3′ end of traA) and the 5′ end of traB (ending 359 bp from the 5′ end of traB) was cloned into the pGEM-T Easy vector (Promega). The PCR insert was then excised via digestion by EcoRI and cloned into the EcoRI site of the suicide vector pGWA34.2 (Table (Table1).1). Since the vector pGWA34.2 does not contain a replication region that works in Bacteroides spp., it must integrate into the host chromosome by homologous recombination to be maintained. Proper placement of the insertion was confirmed by Southern blotting. The vector and traAB segment were then transferred into BT4107 (BT4100 containing CTnDOT) via triparental mating as described previously (26). The transconjugants were then tested for the ability to transfer to BT4001 recipients.
Primer extension was used to determine the transcriptional start sites of rteR and traA. Primer extension analysis was performed with primer extension system-avian myeloblastosis virus reverse transcriptase (Promega). The oligonucleotide primer 5′ CGT ATT GCC ATT CAT ACG CAC GCA 3′ was complementary to nucleotides [nt] 45 to 69 of the suspected traA 5′ leader region. The oligonucleotide primer 5′ GAA CCC TTT GAG CAG CTT GGT ATG CCC TTC G 3′ was complementary to nt 23 to 54 of the possible rteR coding region. Primers were labeled with [γ-32P]dATP (3,000 Ci/mmol; 10 Ci/ml [PerkinElmer]). Total RNAs (60 μg) from tetracycline-induced and from uninduced cells were used for each primer extension reaction. Labeled primers were incubated with RNA at 58°C for 20 min to anneal primers. Avian myeloblastosis virus reverse transcriptase (Promega) was added, and the mixture was incubated for 30 min at 42°C for strand synthesis. Labeled products were run on a 6% denaturing polyacrylamide urea gel. A DNA sequencing ladder was prepared with a template encompassing the transcriptional start site region by using the same radiolabeled primer used for the primer extension reaction. DNA sequencing was done by a Sequenase version 2.0 DNA sequencing kit (U.S. Biochemicals).
Qualitative reverse transcription (RT)-PCR analysis was used to determine whether the genes in the tra operon were members of a single operon and whether the expression of tra genes was controlled at the transcriptional level. Primers were designed to amplify the regions between adjacent tra genes (Table (Table2).2). If two adjacent genes are cotranscribed, an RT-PCR product should be detectable. Primer binding and correct product size were tested initially using chromosomal DNA from BT4007 (Table (Table1).1). BT4007, which contains a single copy of CTnDOT in the chromosome, was grown in the presence (induced) or absence (uninduced) of tetracycline (1 μg/ml). RNA was then isolated from BT4007 via the RNeasy-mini prep kit (Qiagen). Extracted RNA was treated with TURBO DNase (Ambion) to remove any potential DNA contamination. The RNA concentration was then determined by measuring the optical density (OD) at 260 nm. RT-PCR was performed using the Access RT-PCR system (Promega). All RNA preparations were tested in mixtures that did not contain reverse transcriptase to ensure that all DNA had been removed. RT-PCR products were separated on a 1.2% agarose gel.
To quantitate the expression of the tra genes, quantitative RT-PCR (RT-qPCR) was used. BT4007 was grown in the presence (induced) or absence (uninduced) of tetracycline (1 μg/ml). RNA was then isolated from BT4007 via the RNeasy mini-prep kit (Qiagen). Extracted RNA was treated with TURBO DNase (Ambion) to remove any potential DNA contamination. The RNA concentration was then determined by measuring the OD at 260 nm. First-strand synthesis of cDNA was then performed using 1 μg of the extracted RNA, random hexamers (deoxy-N6, where N is any nucleotide; NEB), and Moloney murine leukemia virus reverse transcriptase (NEB). Primers were designed using the Beacon primer design program (Bio-Rad). Primers were made for the traG and sigma 70 genes. The sigma 70 gene was used as an internal control. Five nanograms of each cDNA sample was mixed into Sybr green supermix (Promega), and the samples were placed into a 96-well microtiter PCR plate. The samples were then analyzed via qRT-PCR with an iCYCLER iQ thermocycler (Bio-Rad). RT-qPCR data were analyzed with the iCycler iQ program (Bio-Rad). The results were expressed as the difference (N) between the number of target gene copies and the number of sigma 70 gene copies and were determined by the following equation: N = 2ΔΔCT = 2[ΔCT(traG) − ΔCT(sigma 70)], where ΔΔCT is the ΔCT for induced samples minus the ΔCT for uninduced samples and ΔCT is the difference in threshold cycles for target and reference genes (1, 17).
In order to determine whether the putative ORF in the repressor-containing region, designated rteR, encoded a protein, the possible start codon and the second codon of the ORF were changed to stop codons using the QuikChange II site-directed mutagenesis kit (Stratagene). The template clone for both amino acid mutations was pGFK145.3, which contains approximately 0.78 kb of the rteR region cloned into pCR2.1 (Invitrogen). The double mutant created by site-directed mutagenesis was then cloned into the SmaI to SstI sites of pLYL05 to create pGFK151. The rteR mutant was then transferred into BT4001ΩQABC with pLYL72 by triparental mating as described previously (26).
To determine whether the putative small RNA RteR acted by destabilizing the tra operon transcript, Bacteroides cell cultures (rifampin [rifampicin] sensitive) were grown to an OD (at 650 nm) of 0.7. An initial aliquot of 1 ml was removed from the culture at the point designated time zero. Next, rifampin (final concentration of 100 μg/ml) was added to the cell culture, and further samples (1 ml) were then removed at 15, 30, and 60 min. Samples were placed into RNAlater (Ambion) and placed on ice. RNA was isolated from the samples (by using the RNeasy mini-prep kit; Qiagen) and treated with TURBO DNase (Ambion) to remove possible DNA contamination. cDNA was synthesized using the same method described above for RT-qPCR. Five nanograms of each cDNA sample was used for RT-qPCR. RT-qPCR was performed as described above to determine the amount of traG mRNA. Twenty-five nanograms of each cDNA sample was used for qualitative RT-PCR. GoTaq DNA polymerase (Promega) was used for the final step of qualitative RT-PCR. RT-PCR cycles for each RNA sample (with or without rteR) and primer type (traG or sigma 70 gene primers) were run until a strong band was present on a 1.2% agarose gel. RT-PCR analyses of the 15-, 30-, and 60-min samples were then performed using the time zero RT-PCR program for each given RNA type (with or without rteR) and primer type (traG or sigma 70 primers). Products were separated on a 1.2% agarose gel to determine traG mRNA stability.
Northern blot analyses were performed with membranes containing total RNA. Approximately 6 μg of total RNA was loaded into each lane of a polyacrylamide gel with Tris-boric acid-EDTA buffer. After the gels were run, the RNA was transferred onto a Duralon-UV membrane (Stratagene) overnight (18 mA). After transfer, the membrane was fixed using UV cross-linking. The membrane was kept at −20°C until use. In order to prepare a probe for detecting the target mRNA, a double-stranded DNA fragment containing the target sequence was amplified by PCR. The DNA fragment was labeled with 32P by a random priming method (Promega). The radioactivity of the probe was measured in a scintillation counter, and a quantity of the labeled probe corresponding to approximately 2 × 107 cpm was used for hybridization. The probe was heated to 100°C for 10 min and then quickly placed on ice for 2 min before hybridization. PerfectHyb Plus hybridization buffer (Sigma) was used for hybridization. The membrane was prehybridized with 20 ml of hybridization solution for 30 min at 68°C. The denatured probe was added to the hybridization bottle, and the bottle was incubated at 68°C overnight. After hybridization was complete, the probe was discarded and the membrane was washed four times for 5 min each with low-stringency wash buffer (2× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 0.1% sodium dodecyl sulfate) at room temperature. Then the blot was washed with high-stringency wash buffer (0.5× SSC, 0.1% sodium dodecyl sulfate) at 60°C for 20 min and visualized by autoradiography. The solutions for Northern blot analysis were prepared in 0.1% diethyl pyrocarbonate (Sigma)-treated water (16).
BT4004 cells with and without pGRW3 were grown in the presence and absence of tetracycline. All cells were grown in VPI-maltose (10). DNA was isolated by phenol extraction. PCR analysis of the joined ends of CTnERL was performed. PCR products were separated on a 1.2% agarose gel. Quantitative PCR was performed on the iCYCLER iQ thermocycler (Bio-Rad). RT-qPCR data were analyzed with the iCycler iQ program (Bio-Rad). The results were expressed as the difference (N) between the number of target gene copies and the number of sigma 70 gene copies and were determined by the following equation: N = 2ΔΔCT = 2[ΔCT(JE) − ΔCT(sigma 70)], where ΔΔCT is the ΔCT for BT4004/pGRW3 minus the ΔCT for BT4004 without pGRW3, ΔCT is the difference in threshold cycles for target and reference genes, and JE is joined ends (1, 17).
Previous sequence analysis of the 18-kbp transfer region of CTnDOT revealed that the putative tra genes (traA to traQ) are located in the same 13-kbp region and are all transcribed in the same direction (3). To determine whether the genes from traA to traQ are similarly regulated and may be part of a single operon, RT-PCR was used to detect mRNA that extended from each gene into the adjacent gene. Locations of the amplicons are shown in Fig. Fig.2A.2A. All of the amplification reactions gave RT-PCR products, as expected if the adjacent genes were transcribed as part of a single mRNA. Consistent with this observation, a single-crossover insertion into a segment that contained the 3′ end of traA and the 5′ end of traB eliminated the transfer of CTnDOT to a recipient (yielding <10−8 transconjugants per recipient). This insertion should have left both genes intact but still exerted a polar effect on downstream genes such as traG that are known to be essential for transfer (3). Attempts at Northern analysis were not successful, probably due to the large size (13 kb) of a single transcript.
RT-PCR analyses were also used to determine if the tra operon of CTnDOT was regulated at the transcriptional level. Each tra gene (traA to traQ) was checked for expression in cells exposed or not exposed to tetracycline. Some typical examples of results from these analyses are shown in Fig. Fig.2B.2B. In the absence of tetracycline, transcription was barely detectable or not detectable at all. In contrast, the presence of tetracycline in the medium greatly increased the transcription of all the tra genes. The fact that the pattern of regulation was the same throughout the region further supports our contention that the tra genes are members of a single operon. RT-qPCR was used to quantitate the transcriptional expression of the traG gene. The addition of tetracycline to the B. thetaiotaomicron BT4007 medium caused a 1,293- ± 43-fold increase in traG gene transcription. Thus, increased expression of the transfer genes in cells exposed to tetracycline is regulated at the transcriptional level.
A possible Bacteroides consensus promoter (2) was located 294 bp upstream of the first possible start codon for TraA. Site-directed mutagenesis was used to confirm the location of the putative traA promoter. A traA::uidA translational fusion (Fig. (Fig.1;1; Table Table1)1) was used to monitor promoter strength. A substitution mutation that replaced a T with a G in the putative −7 region of the traA promoter (TTTG → GTTG) reduced GUS activity. The base pair substitution mutant resulted in activity of 25 (±7.9) U/min per mg of protein, in contrast to the wild-type promoter, which resulted in activity of 134 (±29) U/min per mg of protein when induced by tetracycline. Even though the activity was not reduced to the background level by this single base change, the reduction in GUS activity was consistent with the hypothesis that this sequence was within the promoter region.
To confirm the location of the traA promoter, primer extension was used to determine the transcription initiation site (TIS). Results from the primer extension analysis showed that transcription started at a C residue located immediately downstream of the −7 consensus promoter site (Fig. (Fig.3).3). Cells that were grown in the absence of tetracycline lacked the corresponding band in the primer extension assay (Fig. (Fig.3A),3A), as expected if the primer extension product was from the traA message. The distance from the TIS to the first codon of traA was 290 bp (Fig. (Fig.3B3B).
Previously, we showed that RteB activates the expression of rteC and that RteC in turn activates the expression of the orf2c operon (13). To assess the effects of RteC and the excision proteins on tra gene expression, we constructed a system that allowed various combinations of the central regulatory and excision protein genes to be tested for their effects on tra gene expression. In the strains used for these experiments, the central regulatory genes, tetQ, rteA, and rteB or tetQ, rteA, rteB, and rteC, were inserted into the chromosome using the integrase gene and joined ends of mobilizable transposon NBU2 or NBU1 (27). The insertion was designated ΩQAB or ΩQABC, respectively. This inserted DNA cannot be excised because it lacks the NBU1 or NBU2 excision genes (19). Genes from the excision region (pKSO1) and the tra region (pLYL72) were provided on plasmids that have copy numbers of 8 to 10 per cell (Fig. (Fig.11).
Expression of the tra operon was measured initially by RT-PCR using primers that detected traAB, traKL, and traPQ expression, thus monitoring transcription in the beginning, middle, and end of the tra operon. In the presence of both the central regulatory region (ΩQABC) and the excision region (pKSO1), tra operon expression was upregulated in response to tetracycline stimulation. When only ΩQABC was present, however, expression of the transfer genes was detected at a constitutive, intermediate level that was lower than the induced levels of expression yet higher then the uninduced levels of expression detected when the excision region was present. This finding was consistent with our previous observation that pLYL72 transfer was constitutive and occurred at an intermediate level if none of the regulatory genes were present (11). Thus, upregulation of tra gene expression, like the increase in pLYL72 expression above the constitutive level, required both the excision genes on pKSO1 and the regulatory genes that control excision gene expression.
To validate these qualitative results, RT-qPCR was used to measure the relative concentrations of traG mRNA under the different conditions. Without pKSO1, the transfer operon was expressed constitutively, with only a 2 (±<1)-fold difference between mRNA concentrations in the presence and absence of tetracycline. However, when both pKSO1 and ΩQABC were present, the expression of the transfer operon was 549 (±57)-fold higher in the presence of tetracycline than in the absence of tetracycline. The difference between this value and the value of 1,293 (±43)-fold obtained in the case of wild-type CTnDOT (see above) may be due to the fact that in the strain used in this experiment some genes (rteA, rteB, and rteC) were in the chromosome whereas the excision genes and transfer genes were provided on plasmids.
When pKSO1 was present but ΩQABC was absent and there was no tetracycline in the medium, traG expression from pLYL72 was repressed 5 (±<1)-fold compared to the constitutive level of expression seen in the absence of pKSO1 and tetracycline. When pKSO1 was replaced by pKSO5, a plasmid that contains only 500 bp of the 3′ end of the operon and thus provides no excision proteins, expression of traG was not detectable at all. These findings were consistent with our earlier finding that pLYL72 transfer was decreased by 3 orders of magnitude in the presence of pKSO5. This result showed that there is a factor separate from the excision proteins that is repressing tra gene expression below the constitutive level. Our observation that the effect of pKSO1 was less than that of pKSO5 despite the fact that the DNA cloned into pKSO5 is contained within pKSO1 is due probably to basal production of excision proteins in the absence of tetracycline. We had observed previously that although tetracycline stimulation enhances the expression of the excision genes, some expression is still detectable in the absence of tetracycline (14). Thus, basal stimulation of transfer gene expression in the case of pKSO1 is able to counter the effect of the inhibitory factor to some extent. In the case of pKSO5, no expression of excision genes can occur because the promoter and most of the excision gene coding regions are missing.
Having excision proteins act as regulatory proteins is highly unusual, so we wanted to determine whether these proteins were actually controlling the increased expression of the tra operon directly or were acting as accessory factors to facilitate the action of RteC. To do this, we needed to place the orf2c operon under the control of a heterologous promoter. The heterologous promoter we chose was the promoter of susA, a maltose-regulated promoter whose expression is independent of tetracycline (5). To measure the effect of our orf2c construct, pGRW3, we used a translational traA::uidA fusion (pGW40.5) to measure tra gene expression instead of RT-PCR analysis of genes on pLYL72 (Table (Table1;1; Fig. Fig.1).1). We used the fusion construct to monitor the expression of the tra genes because we were unable to introduce the new orf2c construct into the strain carrying pLYL72, probably due to toxicity caused by overproduction of transfer proteins. We were able to introduce pGRW3 into a strain carrying a chromosomal copy of CTnERL, a CTn that is virtually identical to CTnDOT except that it lacks an ermF gene, but this strain grew very slowly, and during the several days and numerous generations required for the transfer assay, the plasmid acquired deletions. In the case of the fusion assay, only a relatively small number of generations were required and pGRW3 did not sustain deletions during this time period.
The GUS activity produced from the traA::uidA fusion on pGW40.5 exhibited the same pattern of upregulation as the RT-PCR results. That is, GUS activity was 17-fold higher in cells exposed to tetracycline than in cells not exposed to tetracycline (134.0 ± 29 versus 7.7 ± 1 U per mg of protein) if both the orf2c operon and ΩQABC were present. The orf2c operon was placed under the control of the susA promoter (PsusA), producing pGRW3 (see Materials and Methods). The growth of BT4001 containing only pGRW3 and pGW40.5 (the tra::uidA fusion) resulted in elevated production of GUS activity, which was independent of tetracycline stimulation. The GUS-specific activity of cells grown on glucose was 919 (±73) U per mg of protein. The GUS-specific activity of cells grown on maltose was 1,989 (±47) U per mg of protein, a level that is about 10-fold higher than that detected in cells carrying the orf2c operon with its natural promoter. The high level of activity in cells grown on glucose is not surprising since previous experience with PsusA has shown that it is a leaky promoter.
We also tested strain BT4004 (BT4001 with a copy of CTnERL) for the effect of pGRW3 on excision mediated by the orf2c operon. To detect excision, PCR was used to detect the joined ends of the circular form (4). This assay, like the assay for activity of the traA::uidA fusion, requires that the cells go through relatively few generations, so instability of pGWR3 was not a problem. In this strain, CTnERL was excised at a higher level when pGRW3 was present. Excision triggered by pGWR3 was independent of tetracycline (Fig. (Fig.4).4). Quantitative PCR analysis revealed that the effect of CTnERL induced by tetracycline was over 1,600-fold greater than the effect of CTnERL in cells grown in the absence of tetracycline but with pGRW3.
pKSO5, a clone that contained a portion of the 3′ end of the exc gene, was all that was needed for inhibition of pLYL72 transfer (27). A small ORF found within this region was tentatively designated rteR. The predicted rteR ORF could encode a very small protein (66 amino acids). To determine whether rteR encoded a protein that is responsible for the observed inhibition of transfer, two different stop codons were introduced into the sites of the putative ATG start codon and the second codon. Transfer efficiency of pLYL72 was then measured with the mutated rteR on pGFK151 provided in trans to determine if inhibition of transfer still occurred. The mutant rteR still inhibited transfer of the pLYL72 plasmid (yielding <10−8 transconjugants per recipient, compared to 10−5 to 10−6 transconjugants per recipient for pLYL72 alone). This result indicated that RteR may not be a protein encoded by the putative ORF in the region but may be a noncoding regulatory RNA. It is still possible that some downstream start codon enables a truncated form of RteR protein to be produced, but if so, the resulting protein would be very small.
To ascertain that rteR was being transcribed and to determine if its expression, like that of the excision genes, was regulated by tetracycline, Northern blot analysis was used to compare the levels of RteR in BT4007 cells (carrying CTnDOT) grown in the presence and absence of tetracycline. Northern blot analysis revealed that a small RNA, RteR, was expressed and was slightly less than 100 nt in size (Fig. (Fig.5A).5A). In contrast to the tetracycline-regulated expression of the excision genes, rteR expression was detected both in the presence and in the absence of tetracycline. The probe should also bind to exc mRNA. In fact, larger transcripts in the lane corresponding to the presence of tetracycline could sometimes be seen in the blot, especially if the autoradiogram was exposed for a longer time period (data not shown).
Primer extension analysis was used to determine the TIS of rteR (Fig. (Fig.5B).5B). This analysis confirmed that RteR was produced constitutively. The 5′ end of rteR was located within the exc gene, starting 1,873 nt downstream of the exc ATG start codon (Fig. (Fig.1).1). Attempts to determine more precisely the 3′ end of RteR by using 3′RACE (rapid amplification of cDNA 3′ ends) failed, so we cannot rule out the possibility that RteR is actually larger than it appears to be on the Northern blot. In fact, a subclone of pKSO5, which removed DNA to within about 50 nt of the apparent end of RteR, as estimated from the Northern blot, no longer inhibited pLYL72 transfer, so RteR may well be larger than 150 nt and the product seen on the Northern blot may have been partially degraded during the RNA isolation process. Further work is needed to define the 3′ end of RteR more precisely.
Small RNAs can cause instability of mRNA by binding to the RNA and attracting the protein Hfq, ultimately leading to the degradation of the RNA (6, 9, 12, 21). Thus, the decreased number of traG transcripts in the presence of RteR may be due to instability of the message rather than transcriptional regulation. To test whether RteR affected the stability of tra mRNA, RNA was isolated from cells for up to an hour after rifampin was added to the medium and the amount of traG message was detected by RT-PCR. Since pKSO5 completely eliminated the production of traG mRNA (see above), pKSO1 was used in this experiment because it permitted lower but still detectable levels of expression of tra genes. Results of RT-PCR analysis are shown in Fig. Fig.6.6. When the repressor RteR was absent, tra mRNA was still detectable after 60 min. RT-qPCR analysis showed that by 60 min after rifampin addition, the abundance of traG mRNA had decreased to about 20% (±3%) of the mRNA detected at time zero. A similar decline was seen in the case of the control sigma 70 mRNA. When RteR was present, approximately the same proportion of traG mRNA found in the absence of RteR was present at 60 min. RT-qPCR analysis showed that the amount of traG mRNA remaining at 60 min was 25% (±7%) of the amount present at time zero. A similar proportion of sigma 70 mRNA was detected after 60 min. Thus, there was no evidence for increased instability in the presence of RteR. Consistent with this result, no Hfq gene homologs have been located in the B. thetaiotaomicron genome sequence.
Our RT-PCR results show that the tra genes are transcriptionally linked. It is possible that more than one promoter is present in the 13-kbp tra region. Even if there is more than one promoter, however, our RT-PCR results show that all promoters respond similarly to tetracycline stimulation and are regulated at the transcriptional level. A translational fusion of traA to the uidA gene showed the same increased expression of traA in cells stimulated by tetracycline, so there was no obvious effect of the fusion on translation. The translational fusion, however, did not exhibit the inhibitory effect of RteR that was detected by RT-qPCR. This result may be due to the insensitivity of the fusion assay relative to the RT-PCR assay.
An important finding was that the same proteins that cooperate with the CTnDOT integrase to excise and circularize the element are also responsible for upregulation of the tra operon, except that in the case of tra gene regulation the integrase is not involved. That is, the excision proteins are not merely accessory proteins that facilitate regulation of the tra genes by RteB or RteC. Orf2c and Orf2d are small basic proteins, and Orf2d has a helix-turn-helix motif. Thus, Orf2c and Orf2d may be DNA binding proteins. Work is under way to determine if and where they bind DNA in the tra gene promoter region. Exc has topoisomerase activity in vitro, but a mutant lacking this activity was able to catalyze excision normally (23), so Exc may also bind DNA.
Our results show clearly that the increase in tra gene transcription activity and pLYL72 transfer frequency seen in cells stimulated with tetracycline is separate from the reduction of transcription and transfer frequency. A small RNA, RteR, appears to be responsible for the negative regulation. The effect of RteR occurs at the transcriptional level but does not involve decreased stability of the tra mRNA. Premature termination forced by the binding of a small RNA (mD) to a message has been seen in the case of another mobile element, pAD1 (24). The effect of RteR on tra gene expression may be similar.
Whatever the mechanism, it is clear that the regulation of CTnDOT excision and transfer is complex. The complex regulatory system may contribute to the stability of the CTn. The need to keep the levels of expression of excision and transfer genes low is evident from the toxic effect of expressing the excision genes at a higher than normal level (by using pGRW3). This toxicity may be due to a combination of the highly increased amount of excision (Fig. (Fig.4)4) and the increased expression of transfer genes.
This work was supported by a grant (AI/GM 22383) from the National Institutes of Health.
We thank Gabrielle Whittle for vectors and preliminary results.
Published ahead of print on 21 August 2009.