INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Bacteroides species harbor integrated mobilizable elements, called mobilizable transposons (MTns). The first step in mobilization of an MTn is excision and circularization, an activity that requires the trans action of two regulatory proteins, RteA and RteB, which are provided by a coresident conjugative transposon (CTn). The circular form of the MTn is then mobilized by other CTn-encoded transfer proteins to a recipient cell, where it integrates into the chromosome (15, 23). An example of an MTn is NBU1, a 10.3-kbp element that is excised and mobilized by CTns of the CTnERL/CTnDOT family. CTnERL and CTnDOT are virtually identical except that CTnDOT carries an additional 13-kbp region (18, 29). Integrated elements that cross-hybridize with NBU1 have been found in over half of all Bacteroides strains tested, representing at least 10 different species (28). Thus, excision and transfer of these MTns appears to occur commonly in nature.

The rteA and rteB genes that are required to initiate excision and transfer of NBU1 are part of an operon that also contains the tetracycline resistance gene tetQ. The proteins encoded by rteA and rteB appear to be members of a two-component regulatory system in which RteA is the sensor and RteB is the response regulator (14). Downstream of the tetQ-rteA-rteB operon is a third regulatory gene, rteC, whose expression is controlled by RteA and RteB. RteC controls many CTn excision and transfer functions but does not appear to be involved in NBU1 excision (1a, 14, 26). Transcription of the tetQ-rteA-rteB operon is induced by tetracycline (10, 25). This explains why exposure of a donor cell to tetracycline is required to trigger excision and transfer of NBU1 (15, 23, 25).

Excision of NBU1 appears to be a complex process involving several NBU-encoded genes as well as the CTn genes rteA and rteB (22). The minimal excision region of NBU1 is indicated in Fig. 1. It includes the integrase gene (intN1), four genes of unknown function (orf2, orf2x, orf3, prmN1), the transfer origin (oriT), and two-thirds of the mobilization gene (mobN1). IntN1, together with the joined ends of NBU1, is necessary and sufficient for integration of the NBU1 circular form. The intN1 gene from either the integrated element or a plasmid is expressed constitutively (21). In contrast to excision and mobilization, integration does not require the trans action of any of the proteins encoded by the CTn. Thus, although intN1 is essential for excision as well as integration (21), it is not responsible for the tetracycline regulation of excision. The MobN1 protein is required for mobilization of the NBU1 circular form (7, 9). MobN1 nicks at the oriT to initiate transfer of the NBU1 circular intermediate through the mating apparatus provided by the CTn. Approximately two-thirds of the mobN1 gene is required for excision (Fig. 1).

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Schematic representation of the integrated NBU1. NBU1 integrates site specifically into the 3' end of a tRNALeu. The chromosomal DNA at the junctions of the integrated NBU1 is indicated by dotted lines and the ends of NBU1 are indicated by hatched boxes. The minimal region of NBU1 required for integration and excision contains the left end, including intN1 through two-thirds of the mobN1 gene as indicated by the arrow, and the right end from the ScaI site (22). Between PvuII and ScaI, just downstream of mobN1 and highlighted in gray, are four orfs of unknown functions that are not necessary for integration or excision. Below the map of NBU1 (accession no. AF238307) is the region containing mobN1, oriT, and prmN1 cloned on the plasmid pPOM. This is the region that is nearly identical to the same region on NBU2 (accession no. AF251288).

To obtain more information about how NBU1 excision is regulated, it was first necessary to ascertain whether expression of any of the NBU1 genes is regulated. We focused our attention on two genes required for NBU1 excision and mobilization, prmN1 and mobN1. There were two reasons for this decision. First, the prmN1-oriT-mobN1 region (Fig. 1) is highly conserved between the two NBUs studied to date, NBU1 and NBU2 (9). Outside this region, there is little sequence similarity between the two NBUs (22, 28). Since both NBU1 and NBU2 require RteA and RteB for excision, it seemed likely that the conserved genes would be the regulated ones if RteB was activating their expression. Second, previous studies had shown that a multicopy plasmid carrying prmN1, oriT, and mobN1 (pPOM; Fig. 1) could completely suppress excision of a wild-type NBU1 when provided in trans with the NBU1 (22). The suppression of NBU1 excision required all of prmN1-oriT and 450 bp of mobN1. Smaller plasmids carrying parts of this region (prmN1-oriT, oriT, or mobN1-oriT) did not have any effect on excision of NBU1. No other DNA segment from the excision region had a suppressive effect (22). Thus, PrmN1 and the N-terminal two-thirds of MobN1 appear to be important proteins in the excision process. In this report, we show that contrary to expectation, both mobN1 and prmN1 are expressed constitutively (not induced by tetracycline) and that what initially appeared to be regulated expression of these genes was due to a time-dependent increase in the copy number of the circular form of NBU1 after excision from the chromosome.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and growth conditions. The bacterial strains and plasmids used in this study are shown in Table 1. Escherichia coli strain DH5MCR (Gibco-BRL) was used as a host for plasmid construction. E. coli strains BL21(DE3) (27) carrying the pLysS inhibitor plasmid (Novagen) and M15 carrying the pREP4 repressor plasmid (Qiagen) were used for bacterial induction of the His₆-tagged proteins. The E. coli strains were grown aerobically in Luria-Bertani (LB) broth or plated on LB agar plates at 37°C. The Bacteroides strains labeled BT (e.g., BT4001 and BT4004) are derivatives of Bacteroides thetaiotaomicron 5482A, also called ATCC 29148 (Virginia Polytechnical Institute Anaerobe Laboratory, Blacksburg, Va.). The Bacteroides strains are grown anaerobically in prereduced trypticase-yeast extract-glucose (TYG) (6) broth or on TYG agar plates incubated in BBL GasPak jars at 37°C. For tetracycline induction of the regulatory operon the Bacteroides strains were grown in TYG broth or on TYG agar containing 2 µg of tetracycline/ml (20).

Vectors and strain construction. The plasmids used for cloning are described in Table 1. For insertional disruptions, the DNA fragments of interest were cloned into either pCQW1 (5) to try to detect transcription or into pGERM (22). These two vectors replicate in and can be mobilized out of E. coli by IncP plasmids but cannot replicate in B. thetaiotaomicron (see Table 1). The resulting vectors were mobilized into BT4104N1-3 (which contained integrated copies of NBU1 and CTnERL), with selection for the erythromycin (10 µg/ml) resistance genes on the plasmids. All insertional disruptions were checked by Southern blotting to confirm that they had inserted in the expected NBU1 location. For complementation tests, the NBU1 segment was cloned into pNLY1, which contains a chloramphenicol (15 µg/ml) resistance marker, or pLYL05 or pLYL7, which carry cfxA, the cefoxitin (20 µg/ml) resistance gene (8, 22). These plasmids are shuttle vectors based on pFD160 (24), which replicates both in E. coli and in B. thetaiotaomicron strains. All of the NBU1 integrative derivatives containing specific deletions were cloned into pGERM and were then transferred by conjugation from E. coli donors to BT4004 recipients, where they site-specifically integrated into the 3' end of a tRNA_Leu gene in the chromosomal target site, attBT1-1 (21). The vectors containing NBU1 regions used for complementation attempts were transferred by conjugation into various Bacteroides strains containing pNBU1 derivatives or insertion mutants in NBU1 and were tested for excision of the NBU1 as previously described (22).

DNA manipulations. Plasmid DNA was isolated from E. coli and Bacteroides strains by the method of Ish-Horowitz as described by Sambrook et al. (16). Restriction digests and ligations were performed essentially as specified by the manufacturer (Bethesda Research Laboratories, Inc., Gaithersburg, Md., or New England Biolabs, Inc., Beverly, Mass.). Chromosomal or total DNA preparations were made using a modification of the method of Saito and Miura (13) as previously described (22).

Southern blots. NBU1 excision assays were conducted as described by Shoemaker et al. (22). Most of the insertional disruptions and deletions in NBU1 genes have been previously described and most were recloned into pGERM for this study (22). The newly constructed deletions in orf2 and orf2-orf3 are described in Table 1. The 1.7-kbp HincII fragment of NBU1 containing the joined ends or the attN1 region was used as the probe (21). The fragment used for mobN1 detection to follow increased copy number of the NBU1 in a tetracycline-induced culture was the 1-kbp HincII-PvuII 3' end of mobN1. The 0.9-kbp 5' end of the chromosomal starch utilization gene susG (17) was used to probe for a single copy gene in the same strain. The probes were labeled with fluorescein-dUTP by using random primers as specified in the NEN Life Sciences Renaissance kit protocol. The Southern blots were developed using a chemiluminescent substrate and exposure to film.

Overexpression and purification of mobN1 and prmN1 gene products in E. coli. The pET T7 promoter expression system (Novagen) was used to overexpress N-terminally His₆-tagged MobN1 (27). The mobN1 gene was PCR amplified from B. thetaiotaomicron strain BT4104N1-3, which contains an integrated NBU1, with primers (mobN1-F [bp 8584], AA G AAT TCA TGG CAA CAA AAT CAA BCA TAC AC, and mobN1-R [bp 10,011], TC G AAT TCT ATC ATA ATT ACA TTC TGA ATC CT) that contained EcoRI sites (shown in bold letters). The PCR product was first cloned into PCR product cloning vector pGEM-T (Promega) and the product was sequenced to confirm that there were no mutations. A 2.4-kbp EcoRI fragment was then isolated and cloned into expression vector pET-28a(+) digested with EcoRI. The resulting clone, pETmobN1, was used to transform the expression host strain BL21(DE3) carrying the inhibitor plasmid pLysS. Kanamycin-resistant transformants (Kn^r, 50 µg/ml) were selected. The T5 promoter-based QIAexpress system (Qiagen) was used to overexpress C-terminally His₆-tagged PrmN1. The full-length 971-bp prmN1 was PCR amplified from B. thetaiotaomicron BT4104N1-3 with the forward primer containing an NcoI site (shown in bold) (prmN1-F [bp 7420], CGC AAG A CC ATG GCA ATA GAA GAA) and the reverse primer containing a BglII site (shown in bold) (prmN1-R [bp 8391], AGT GGG AGA TCT CCG AAA GCC GTT TTT). The PCR product was digested with NcoI and BglII and inserted directly into the corresponding sites in pQE60. The resulting clone, pQE60prmN1, was used to transform the expression host strain M15 carrying repressor plasmid pREP4.

Generating polyclonal antisera against MobN1 and PrmN1. The proteins were purified from acrylamide gels and the purified MobN1 and PrmN1 were dialyzed using a microdialyzer (PGC Scientifics). Approximately 1 mg of each recombinant protein was used to immunize mice. Ascites fluid was collected, delipified, and stored in saturated ammonium sulfate until needed. Antibodies were generated by the Immunological Research Center of the University of Illinois.

Membrane preparations. Both MobN1 and PrmN1 were found primarily in the membrane fraction of the sonicated cells. Therefore, unless otherwise noted, the membrane fractions were the source of proteins for all experiments. To obtain the membrane fractions, B. thetaiotaomicron strains containing various NBU1 derivatives or plasmids were grown anaerobically in 100 ml of prereduced TYG broth to late exponential phase (OD at 650 nm of 0.7 to 0.9) with or without tetracycline at 2 µg/ml. The cells were harvested at 4°C (10,000 × g for 20 min). The cell pellet was washed once with 20 mM potassium phosphate buffer (pH 7.2) and resuspended in 5 ml of the same buffer, and cells were disrupted by sonication. The sonicated samples were spun at 10,000 × g at 4°C for 15 min to remove large cell debris. The membranes were pelleted from the supernatant by ultracentrifugation (200,000 × g for 2.5 h at 4°C). The soluble fraction was collected and saved, and the membrane pellet was thoroughly resuspended in 7 ml of the 20 mM phosphate buffer and pelleted again by ultracentrifugation under the same conditions. The membrane pellet was resuspended in 200 to 300 µl of 20 mM potassium phosphate buffer, and the membranes were dispersed by gentle sonication for 15 s. For the soluble fraction, the supernatant from the first ultracentrifugation step was concentrated using an Amicon Centriprep YM-10 system to about 0.5 ml. If no membrane-associated proteins were detected, the soluble fractions were checked in case there was a localization effect due to the conditions used or the NBU1 derivative being tested. The concentrations of the soluble and membrane proteins were determined using a Bio-Rad D_C Protein Assay kit.

Immunoblotting of MobN1 and PrmN1. To detect the expression of MobN1 and PrmN1, membrane preparations were made from B. thetaiotaomicron BT4104N1-3 containing CTnERL and a wild-type NBU1 and from other BT strains containing derivatives of NBU1 (Table 1). The membrane fraction was used because most of the MobN1 and PrmN1 protein fractionated with the membranes. Using a membrane fraction made the assay more sensitive. Samples containing 100 µg of protein were loaded in each well. The proteins were separated on sodium dodecyl sulfate-10% polyacrylamide gels. Proteins were electrotransferred from the gels to a nitrocellulose membrane using a Bio-Rad Trans-Blot. The membranes were incubated with polyclonal antibodies, usually diluted 500- to 1000-fold in TTBS (20 mM Tris [pH 7.5], 0.2% Tween 20, 0.5 M NaCl) with 1% bovine serum albumin, against MobN1 and/or PrmN1 at room temperature for 3 h to overnight. MobN1 and PrmN1 were detected by using a Bio-Rad goat anti-mouse horseradish peroxidase Opti-4CN kit and the membranes were scanned or photographed. The truncated form of MobN1, calculated to be about 30.6 kDa, could not be detected in either the membrane or the concentrated soluble protein preparations. In cases where no MobN1 or PrmN1 was detected in the membrane fractions, the soluble fraction from the same cells was concentrated to the same degree as the membrane fraction and probed to determine whether the proteins were truly missing or were fractionating aberrantly.

Time course for excision of NBU1 and detection of MobN1. Cultures were grown in 100-ml bottles containing TYG medium with 2 µg of tetracycline/ml added. The bottles were inoculated with 3 ml of overnight cultures of BT4104N1-3 or BT4004(pPOM) grown without tetracycline. Five milliliters of each overnight culture was used to make total DNA preparations and the remainder of the culture was used for total membrane preparations. The tetracycline-induced cultures were incubated at 37°C and the OD at 650 nm was monitored. Samples were taken at ODs of 0.3, 0.5, and 0.7 and overnight (ca. 0.8 to 0.9), and the times required to reach these densities were noted. At each growth point, 10 to 20 ml of the cultures was used to obtain DNA for the excision assays by Southern blots and the remaining 90 ml (180 ml for the samples with an OD of 0.3) was used to obtain membranes and soluble fractions for the Western blot analysis. The estimate of the copy number of NBU1 at the various time points was determined by analyzing the Southern blots of the samples probed with the mobN1 and susG probes (see above). Serial dilutions of the various samples from different time points were included on the Southerns to get the intensity of the bands in the linear range of the film. The total density of each band was determined using the Bio-Rad gel documentation and Quantity One Software analysis system. The ratio of the total density of the mobN1 bands to that of the susG bands for the uninduced BT4104N1-3 sample was normalized to 1 and the ratios obtained for all of the other samples were divided by the same factor.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Location of promoter regions of NBU1 excision genes. Previously, we had shown that there was a promoter upstream of intN1 and one upstream of mobN1 (9, 22), but it was not known whether the other NBU1 genes involved in excision were organized in an operon. To answer this question, we first tested cloned regions of the NBU1 excision region for the ability to complement excision-deficient mutants to excision proficiency. None of the clones tested complemented any of the mutants. Accordingly, we turned to the alternative strategy illustrated in Fig. 2A. Single crossover disruptions were made that separated each gene from the one upstream but kept both genes intact. If the two genes that were separated from each other were in the same operon, the insertion should have a polar effect on the downstream gene. Mutant NBU1s with insertions that separated orf2 from orf3 (bp 5147 to 6028) or orf3 from prmN1 (bp 7030 to 7699) were still able to excise as well as the wild type (Fig. 2B). Thus, orf3 and prmN1 are in separate transcriptional units (Fig. 2A, panel 3). Separating intN1 from orf2 (bp 4063 to 5154) abolished excision (Fig. 2B), indicating that intN1 and orf2 are in the same transcriptional unit. Information about the location of promoters was used in subsequent cloning experiments designed to monitor gene expression of prmN1 and mobN1.

View larger version (31K): [in this window] [in a new window]   FIG. 2.   (A) Strategy for determining the approximate location of promoters for various excision genes. The regions containing the 3' end of the upstream gene and the 5' end of the downstream gene (panel 1) were cloned into pGERM, a vector that does not replicate in Bacteroides strains (panel 2). When the resultant suicide vectors were transferred into BT4104N1-3, they integrated by homologous recombination into NBU1. The integrated vectors separated the two targeted genes by about 5 kbp of vector DNA. If the two genes are in an operon, the insertion will have a polar effect on the downstream gene and will eliminate NBU1 excision. Panel 3 shows a summary of the approximate location of the promoters of genes in the minimal excision region. (B) Southern blots showing the results of the excision assay on the various insertions described in panel A. For panel 1, the DNA from each strain was digested with HincII and the blot was probed with the 1.7-kbp HincII DNA fragment that contains the joined ends of the NBU1 circular form. The designation pP2x-3 at the top of the blot indicates the insertion that separated orf2x-orf3 from orf2. Similarly, the other insertions are labeled with the gene that is downstream from the region where the integration occurred. The arrow to the left of the blots indicates the position of the 1.7-kbp joined end fragment that is diagnostic for the NBU1 circular form. JL and JR are, respectively, the left and right ends of the copies of NBU1 that are still integrated in the chromosomal site. For panel 2, the DNA was digested with HincII and KpnI because the intN1'-orf2' fragment used for the pP2N1 insertion had a small region of identity to the probe. This combination of enzymes would give a clearer separation of an excision product (indicated by arrow) from the junction- and vector-containing bands.

Production of PrmN1 and MobN1 from an integrated copy of NBU1. We had found previously that trancriptional fusions of the upstream regions of prmN1 or mobN1 to uidA did not give detectable -glucuronidase activity (21). Thus, we used antibodies against PrmN1 and MobN1 to detect production of the two proteins. Both proteins were detectable on Western blots of extracts from B. thetaiotaomicron carrying CTnERL and NBU1. The proteins were only detectable in cells that had been stimulated by tetracycline (Fig. 3). Moreover, a plasmid carrying the tetQ-rteA-rteB operon, pQAB (pNBS4), was able to replace CTnERL, whereas a plasmid carrying only tetQ-rteA, pQA (pNBS2), was not sufficient. There was also no requirement for rteC, which is carried on pAMS9.

View larger version (38K): [in this window] [in a new window]   FIG. 3.   Western blot showing the pattern of production of MobN1 and PrmN1 from integrated NBU1 and from a plasmid containing the two genes. Membrane preparations for the indicated strains were loaded on a sodium dodecyl sulfate-10% acrylamide gel. The blot containing the transferred proteins was probed with antibodies to MobN1 and to PrmN1. The samples are labeled according to the source of the RteA and RteB proteins in the host strain. The strains were grown with (+) or without () 2 µg of tetracycline (Tc)/ml. The strain labeled CTn contains CTnERL and NBU1 (BT4104N1-3). All of the other strains were derivatives of BT4100N1S1 (NBU1, no CTn) which contained a plasmid with the indicated CTn regulatory genes: tetQ-rteA on pQA (pNBS2), tetQ-rteA-rteB on pQAB (pNBS4), and tetQ-rteA-rteB and rteC on pQABC (pAMS9). pPOM is a plasmid with a copy number of about 5 per cell that contains prmN1 and mobN1 and is in a strain containing CTnERL which provides RteA and RteB when grown in tetracycline. Locations of MobN1 (54 kDa) and PrmN1 (36.4 kDa) are indicated by arrows at the right.

Gene expression from genes cloned on a plasmid. Although expression of prmN1 and mobN1 appeared to be regulated when produced from NBU1, Western blot analysis of proteins produced from a plasmid, pPOM, that carried only prmN1, oriT, and mobN1 revealed that PrmN1 and MobN1 were now produced independently of RteAB and of tetracycline induction (Fig. 3). That is, the proteins were produced at the same level in a strain that contained only pPOM and had not been exposed to tetracycline as in a strain that carried both pPOM and CTnERL and had been exposed to tetracycline. Also, PrmN1 was produced from a plasmid that contained only prmN1 (pPrm-oriT) and MobN1 was produced from a plasmid that contained only mobN1 (pLYL20) (data not shown). These results suggested two possible hypotheses. The first hypothesis was that expression of prmN1 and mobN1 is regulated at the transcriptional level, but expression of the genes on pPOM was constitutive because pPOM did not contain a repressor gene that normally controls these genes. A second hypothesis was that excision and circularization of NBU1 had to occur before there was detectable production of PrmN1 and MobN1.

View larger version (28K): [in this window] [in a new window]   FIG. 4.   (A) Integrated deletion derivatives of NBU1 and NBU1 plasmid clones tested for production of PrmN1 and MobN1. At the top is a schematic of the integrated NBU1 with the open reading frames not necessary for excision between PvuII and SphI indicated by "//." The derivatives of NBU1 with the plasmid (p) designation were replicated in E. coli and integrated in B. thetaiotaomicron. An asterisk by the name indicates that this construct has the entire mobN1 gene. The other constructs have only two-thirds of mobN1 (to the arrow indicating the end of the minimal excising region and depicted by a vertical dotted line) and do not produce any MobN1 detectable with our antibody. Several of the constructs were made to contain the minimal excising region and as a result they are missing the region downstream of mobN1, extending to the SphI site near the right end of NBU1. The deletions downstream of the end of the minimal region are shown as gray filled rectangles. The open boxes indicate the deletions in the genes within the region essential for excision. At the bottom of the figure are the plasmids used to detect the production of PrmN1 and/or MobN1. pY5 is the replicative form of the Y5D shown in the top section. To the right of the drawing are two columns summarizing the results of Western blot detection of MobN1 and PrmN1. +C, the protein was produced constitutively and did not require induction of the cells with tetracycline. (B) Western blots on which the results listed as "+" and "" for some of the integrated derivatives of NBU1 shown in panel A are based. Membrane proteins (100 µg) were loaded for each sample. The Western blot was developed for both MobN1 and PrmN1 and the expected position for each of the two proteins is indicated on the right. The predicted location for the 30.6-kDa truncated form of MobN1 (MobN1') is indicated by an arrow near the bottom of the blot. This truncated form of MobN1, if produced, was not detectable by our antibodies in total membranes or in soluble fractions (not shown). Tc, cultures grown without tetracycline induction.

Excision is required for detectable production of PrmN1 and MobN1. It was noteworthy that the plasmid that contained all of NBU1 (pY5; Fig. 4A) produced detectable levels of both PrmN1 and MobN1 and production was independent of tetracycline, similar to that observed for pPOM (data not shown). However, the integrated form of pY5 (pY5D) produced no detectable PrmN1 or MobN1 (Fig. 4B). Moreover, any mutation in integrated NBU1 that blocked excision was associated with no production of PrmN1 and MobN1. These results supported the hypothesis that the amount of the excised circular form might be important for the production of MobN1 and PrmN1. If, after excision, the circular form of NBU1 reached a high enough copy number, production of PrmN1 and MobN1, which was not detectable when the NBU1 was present only in a single copy, would become detectable. That is, genes that were actually expressed constitutively on integrated NBU1 would appear to have tetracycline-induced expression because after NBU1 excision, copies of the NBU1 circular form would accumulate to a high enough number for proteins produced at their basal level to become detectable.

View larger version (60K): [in this window] [in a new window]   FIG. 5.   Correlation of the appearance of MobN1 and the appearance of the excised circular form of NBU1 following induction with tetracycline. (A) The top half of the panel is a Southern blot showing the assay for the joined ends of the excised circular form of NBU1. The DNA from the culture grown to the indicated OD following tetracycline (Tc) induction was digested with HincII and the blot was probed with the 1.7-kbp HincII joined-ends fragment of NBU1. The lower strip is a Western blot of membrane preparations from cells taken at the same time point as for the Southern blot (see Materials and Methods). The lanes are labeled with the OD of the culture from which the DNA and total membrane proteins were obtained. (B) Results shown here are from a duplicate experiment of the one shown in panel A. It is provided to show that there was some variation in the OD at which the circular form of NBU1 and MobN1 appeared. Also, in this experiment, a strain containing pPOM was also included in parallel. As can be seen from the second Western blot strip at the bottom of the panel, MobN1 from pPOM was expressed at all time points in this strain. (C) Growth curve for the samples shown in panel A, with arrows indicating the points at which the samples were taken. (D) The Southern blot shown in panel B was stripped and reprobed with a probe that detects mobN1 and thus the concentration of NBU1 DNA (integrated or circular form) in the cell and a probe that detects an unrelated chromosomal gene, susG. This panel shows that the concentration of NBU1 DNA increases at the time when the circular form of NBU1 appears and accumulates in stationary phase. Copy numbers for mobN1 relative to susG were determined by band density determinations for both NBU1 and pPOM for the same samples shown in panels A and B on a separate Southern blot (not shown).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

When NBU1 or NBU2 is introduced by conjugation into B. thetaiotaomicron, it invariably integrates site specifically into the recipient's chromosome. Because of this, it had been assumed that NBUs do not replicate in B. thetaiotaomicron. In fact, "NBU" stands for "nonreplicating Bacteroides unit." The results of experiments reported here suggest that this view of NBUs may need to be modified. Results reported here show that after excision, copies of the NBU circular form accumulate to an estimated copy number of 2.5 to 2.7 per cell. Whether this is due to transient replication or to an excision method that produces more than one circle remains to be determined. The MobN1 and PrmN1 proteins are only produced at detectable levels at the point when excision begins to occur. When the prmN1 and mobN1 genes were cloned onto a plasmid having a copy number of 5 per cell such as for pPOM (Fig. 5), expression of the genes was independent of growth phase and tetracycline stimulation. Yet during excision of NBU1, expression of these genes appeared to be regulated both by growth phase and by tetracycline. The topology of DNA molecules has been shown to influence transcription of genes (2-4). Possibly, transient changes in the supercoiling of the circular form of NBU1 that occur during the excision process could trigger increased expression of NBU1 genes.

Although excision seems to be required to cause maximal expression of prmN1 and mobN1, the concentration of protein is not completely correlated with the amount of excised circular form. As seen in Fig. 5A and B, the level of MobN1 rises when the circular form first begins to reach detectable levels, and then the amount of MobN1 decreases as the cells enter stationary phase, even though the amount of NBU1 circular form increases or remains constant. This may be due to instability of the MobN1 or PrmN1 proteins or to a decrease in protein synthesis as cells enter stationary phase, because when these proteins are produced from the genes carried on a plasmid (pPOM), the level of the proteins in the cell remains constant in the exponential growth phases but decreases relative to the plasmid copy number in the stationary phase (Fig. 5B).

The increase in copy number of the circular form of NBU1 could be explained if excision involves more than one step: one step produces the excised circular form, and then in a second step the circular form accumulates so there is more than one copy per cell. Having multiple circular forms per cell could increase the probability that the circular form would be nicked by MobN1 and transferred by conjugation. According to this hypothesis, there is a burst of PrmN1 and MobN1 production during the first step but expression of these genes ceases or decreases during the second step as the cells go into stationary phase. The need for increased MobN1 after excision is clear; it is required for mobilization of the circular form from the donor to a recipient. This explanation does not apply to PrmN1, which appears not to be necessary for mobilization (7). PrmN1 might contribute, however, to the step in which the number of copies of the circular form increases. Indirect evidence supporting this hypothesis is that a mutation in prmN1 (pY11D) that abolished excision as detected by Southern blot still allowed some low-level mobilization of the integrated NBU1 to a recipient similar to that observed for pY5D in orf3 (22, 23). This suggests that some excision was occurring even though accumulation of the circular form to the level necessary for detection on a Southern blot was not occurring.

The accumulation of multiple copies of the circular form may not be true replication, in the sense that it involves a plasmid-like replication origin, but could be due to a replicative mode of excision. In the past, it was assumed that NBU1 excision is not a replicative process because it is possible to detect the site from which NBU1 had excised by PCR amplification. Further work may reveal that NBU1 excision does in fact occur by a replicative mechanism under some conditions.

Another surprising finding was that none of the NBU1 excision genes tested was able to complement excision-deficient mutants in trans. From the insertional disruption data and from Western blot data, it is clear that the cloned segments contained a promoter region, and in the case of prmN1 and mobN1, were producing the protein encoded by the gene. In the case of mobN1, we had shown previously that the gene could complement mobilization in trans. The fact that it could not complement a mutant of NBU1 to excision proficiency in trans suggests that for excision the genes are cis acting.

Conjugal transfer of DNA is a form of replication because the transferred single-stranded copy of the element and the single-stranded copy that remains behind are both copied to recreate the double-stranded circular form. This is not the explanation for the increase in copy number of the NBU1 circular form, because the phenomenon occurs in cells that have only the tetQ-rteA-rteB portion of the CTn but not the region that contains transfer genes. If, as our hypothesis suggests, excision occurs in two phases, with the excision event followed by a rise in copy number of the circular form, there should be one or more NBU1 genes involved in the second stage. As mentioned previously, prmN1 could be such a gene. The existence of a second phase in the excision process could explain why so many genes are required for excision. Most excision systems consist of an integrase gene and one other gene. This is true of the lambdoid phages (1), the gram-positive CTn Tn916 (11, 12), and the Bacteroides CTn CTnDOT (1a). Additional proteins may be involved, but these are host factors such as IHF that aid in bending DNA to form the synaptic excision complex. At this point, we suggest an alternative hypothesis, that NBU1 excision can occur by a conservative process under some conditions and by a replicative process under other conditions. Further studies will be needed to determine which hypothesis is correct.