Phage L5 integrating vectors are present within the Mycobacterial Cell in an equilibrium between integrated and excised states

Abstract

Integrating mycobacterial plasmids containing the phage L5 attachment site (attP) are able to insert into the mycobacterial chromosome attB site. Plasmids containing the attP site and chromosome containing the attB site are present in equilibrium between the inserted and the excised states in the presence of the phage L5 integrase.

I. Introduction

Vector systems have been developed to integrate into the mycobacterial chromosome (Lee et al, 1991; Snapper et al, 1988). These systems utilize the mycobacterial phage L5 attachment site attP and its target bacterial attachment site attB in conjunction with the phage L5 integrase (Lee and Hatfull, 1993; Pedulla et al, 1996, Pena et al, 1996, 1997, 1998, 1999, 2000). Integrase thus catalyzes the integration of vectors bearing an attP site into the bacterial attB site. These integrative vectors, however, can be unstable within the mycobacterium and can result in plasmid loss (Lewis and Hatfull, 2000; Springer at al, 2001). It has subsequently been determined that elimination of the integrase gene after integration of a target vector resulted in an extremely stable integrative vector within the mycobacterial cell (Springer et al, 2001). Additionally, transformation of mycobacteria containing a previously integrated vector with an additional L5 integrating vector results in replacement of the first vector with the second vector in the presence of integrase (Pashley and Parish, 2003). Thus integrase mediates L5 vector instability as well as integrated vector exchange. It is possible to insert two integrative plasmids sequentially into the mycobacterial chromosome. When an integrative vector containing an additional plasmid borne attB site is integrated into the mycobacterial chromosome, an additional L5 integrating vector can be inserted into this additional attB site. In the presence of integrase, these two vectors are extremely unstable unless the integrase gene is eliminated (Saviola and Bishai, 2004). I demonstrate that these vectors can be excised from the mycobacterial chromosome and be reintegrated into the mycobacterial chromosome. In a two-plasmid system, it is feasible to observe integration, excision, and reintegration of a phage L5 integrating plasmid. It is shown that integrating vectors are present in equilibrium between integrated and excised states within the mycobacterium. In addition, it is possible to recover L5 integrating plasmid DNA from mycobacterial strains containing a previously integrated vector by this equilibrium.

II. Material and Methods

A. Bacterial strains and plasmids

Mycobacterium smegmatis strain mc² 155 (American Type Culture Collection ATCC, Manassas VA) and Escherichia coli strain DH5α were used for all experiments. Integrating vectors used were pBS11 and pBS29 which contain the phage attP site and the phage L5 integrase as previously described (Saviola and Bishai, 2004). Additionally, the pBS29 contains a bacterial attachment site attB, the pBS11 a hygromycin resistance gene, and the pBS29 a kanamycin resistance gene (Saviola and Bishai, 2004). Vectors pBS33 and pBS37 lacking L5 integrase were used as previously described (Saviola and Bishai, 2004). The pBS33 and pBS37 were created by cutting out the integrase containing PstI fragment of pBS11 and pBS29 respectively. The pBluescriptint contains the integrase gene within a vector which is nonreplicative in mycobacteria (Springer et al, 2003).

B. Vector loss assay

M. smegmatis was transformed with the integrating vectors pBS29 or pBS37 + pBluescriptint and transformants were recovered on 7H10 agar (Difco) + 10% ADC (5% w/v bovine serum albumen Fisher, 2% w/v dextrose EM Science, 0.85% w/v NaCl EM Science) plates containing 50 µg/ml kanamycin. Electrocompetant cells were prepared using the resultant M. smegmatis integrating vector transformed strains. Electrocompetant cells containing integrated pBS29 or pBS37 were then transformed with pBS11 or pBS33 + pBlueScriptint respectively. Transformants were grown overnight in 7H9 broth (Difco) supplemented with 10% ADC in the absence of antibiotic (20 hrs), diluted, and plated on to 7H10 agar (Difco) supplemented with 10% ADC without any antibiotics. Resultant individual single colonies were numbered and then patched onto each of 7H10 agar plates containing either hygromycin 100 µg/ml or kanamycin 50 µg/ml to determine antibiotic resistance. One hundred individual bacterial colonies each were tested for M. smegmatis transformed with pBS29 + pBS11, or pBS33 + pBS37. Kanamycin sensitive strains were assayed for loss of aph via the polymerase chain reaction (PCR). Briefly, M. smegmatis strains were lysed by an initial exposure to 99°C for 10 min, followed 1 min at 95°C, and subsequently primers BS27 and BS28 which bind to the 5’ and 3’ ends of the aph gene within integrated vector DNA were annealed at 55°C for 2 min. Elongation of primers with Taq DNA polymerase (New England Biolabs) was performed at 72°C for 1 min. PCR was performed for 30 cycles excluding the initial 10 min incubation at 99°C on an Eppendorf Mastercycler Personal thermocyler.

C. Plasmid DNA recovery and analysis

Total DNA containing small quantities of excised integrative L5 plasmid was prepared by growing the mycobacterial strains overnight to saturation. Three milliliters of mycobacteria were centrifuged and resuspended in 550 µl 10mM tris (Fisher) and 1mM EDTA (EM Science) (TE) with 10 mg/ml lysozyme (Fisher) and incubated overnight at 37°C. After overnight incubation 70 µl of 10 % SDS (Fisher) and 6 µl 10 mg/ml proteinaseK (New England Biolabs) were added and the reaction incubated at 65°C for 10 min. One hundred µl 5 M NaCl (EM Science) and 80 µl 10 % CTAB (cetyl trimethyl ammonium bromide) and 0.7 N NaCl (EM Science) were added and the samples were further incubated for 10 min at 65°C. The samples were then extracted with phenol chloroform (Fisher) twice and then extracted once with chloroform isoamylalcohol (EM Science) (24:1). DNA was precipitated with 700 µl isopropanol (Acros) for 30 min at −80°C. The samples were centrifuged at 14,000 RPM for 15 min. The supernatants were removed and the pellets were washed with 70 % ETOH (Shelton Scientific). The pellets were then dried and resuspended in TE with RNAse I_f (NEB). Total mycobacterial DNA was quantitated by determining the optical density (OD) 260 nm. Purity of the DNA was estimated by comparison of the ratio of the OD 260 nm to the OD 280 nm on a Genesys 10 UV spectrophotometer (thermospectronic).

Recovery of excised L5 integrative plasmids was performed by transformation of Electocompetant DH5α E. coli with total DNA from M. smegmatis. This total DNA may contain a small amount of excised L5 integrative plasmid. Antibiotic resistant E. coli transformants were recovered on Luria Broth (LB) agar (Difco) containing either 50 µg/ml of kanamycin or 200 µg/ml hygromycin. Plasmid DNA was prepared from these E. coli transformants using the Wizard Miniprep system (Promega). Plasmids recovered from kanamycin and hygromycin resistant E. coli as well as pBS29 and pBS11 were cut with Sau3A1 (NEB) and electrophoriesed on a 2.0 % agarose gel and were compared with each other. Recovered plasmids bearing similarity to pBS11 or pBS29 were retransformed into electrocompetant M. smegmatis and plated onto 7H10 agar plates containing 100 µg/ml hygromycin (Calbiochem) or 50 µg/ml kanamycin (Fisher) respectively.

III. Results

When total DNA is purified from M. smegmatis containing the L5 integrative plasmid pBS11 inserted into its chromosome, the unintegrated form of the plasmid can be recovered in small quantities. Total DNA was prepared from M. smegmatis with integrated pBS11 grown 20 hrs in the absence of antibiotic. Recovery of unintegrated plasmid was performed by transformation of the total DNA into electrocompetant E. coli cells and E. coli bearing the plasmid were selected on 200 µg/ml hygromycin. The rate of recovery of transformants was 0.001 hygromycin resistant transformants/ng of total DNA. Hygromycin resistant plasmid could then be recovered from transformed E. coli by miniprep procedure. All plasmids recovered in this manner could be retransformed into electrocompetant M. smegmatis cells indicating that they are pBS11. This finding indicates that L5 integrative plasmids are present within some of the bacteria in an excised form at least temporarily. Integrated plasmids are likely present in equilibrium between excised and integrated states within host bacterial cell. If the excision rate is slow and the integration rate is rapid this will favor plasmid retention within the mycobacterial cell.

To test the stability of pBS11 in the chromosome, M. smegmatis with pBS11 integrated into the chromosome was incubated overnight in the absence of antibiotic. The mycobacterial culture was then serially diluted and plated onto 7H10 agar media in the absence of antibiotic to recover single colonies. After growth on 7H10 agar plates single colonies were patched onto agar plates containing hygromycin. One hundred single colonies were tested and all grew when patched onto hygromycin agar plates indicating that pBS11 is not readily lost from the host bacterial cell.

The above observations indicate that phage L5 integrating plasmids are in equilibrium between integrated and excised state within the mycobacterial cell, though the plasmids are rarely lost from the bacterium. In order to observe experimentally the excision and the reintegration of pBS11, I used a two-plasmid system previously described (Saviola and Bishai, 2004). This two-plasmid system utilizes an integrating plasmid pBS29 that contains an additional bacterial attachment attB site as well as a kanamycin resistance gene, aph (Figure 1). By integrating into the M. smegmatis chromosome pBS29 will retain an additional attB site and the pBS11 can be integrated into the retained site. M. smegmatis was first transformed with pBS29, competent cells were prepared, and these cells were then transformed with pBS11 and selected for on hygromycin and kanamycin 7H10 plates. pBS29 + pBS11 transformants were grown up overnight for 20 hrs in the absence of antibiotics and plated on 7H10 agar plates in the absence of antibiotics. Individual colonies were numbered and patched on both agar plates containing either kanamycin or hygromycin. The loss rate for kanamycin resistance was 33% while the loss rate for hygromycin resistance was 3%. No colony was found to be sensitive to kanamycin and hygromycin both. Seventeen kanamycin sensitive strains were tested for loss of the kanamycin resistance gene. All strains lost the gene as assayed by PCR, whereas 2 control strains that retained kanamycin resistance maintained the gene as assayed by PCR (Figure 2). Thus mutation of the kanamycin gene is an unlikely explanation of kanamycin sensitivity. Loss of pBS29 containing kanamycin resistance could explain the resultant kanamycin sensitivity. One mechanism to explain kanamycin sensitivity and retained hygromycin resistance within the mycobacterial cell is the excision of pBS11 followed by excision of pBS29 (Figure 3) and reintegration of pBS11 (Figure 4). The pBS11 lacking the additional attB site would effectively block reintegration into the chromosome by pBS29 due to the absence of an available attB attachment site (Figure 4). Thus it is possible to observe excision of pBS11 out of the mycobacterial chromosome, followed by reintegration as evidenced by the loss of pBS29. It is also possible that pBS29 (with pBS11 integrated within pBS29) excises and pBS11 subsequently excises from this extra-chromosomal composite vector. In the previous scenario, pBS11 could reintegrate to block future reintegration of pBS29 (Figure 4). If pBS29 reintegrates first, however, pBS11 can then reintegrate into the retained attB site. Thus as pBS29 lacks the ability to block reintegration of pBS11, this most likely accounts for the low rate of loss of this second plasmid. Rarely pBS11 can fail to reintegrate and thus may be lost from the mycobacterial cell as 3% of bacteria are hygromycin sensitive after growth in the absence of antibiotic for 20 hours. Even more rarely both pBS29 and pBS11 fail to reintegrate and both plasmids are lost from the host cell. No mycobacteria assayed developed this pattern of sensitivity when grown in the absence of antibiotic. Based on the above data pBS11 both excises within the mycobacterial cell and reintegrates within the chromosome.

Figure 1.

pBS29 containing an attB site is integrated into the M. smegmatis chromosome. The additional attB is preserved and can accept integration by pBS11. The resultant M. smegmatis strain is kanamycin and hygromycin resistant.

Figure 2.

M. smegmatis strains lacking kanamycin resistance were assayed by PCR for the presence of aph. One M. smegmatis strain that retained kanamycin resistance revealed the presence of aph as assayed by amplification. Another kanamycin resistant strain was assayed and the aph gene was also amplified (data not shown). Four kanamycin sensitive bacterial colonies lacked amplification of the aph gene. The remaining 13 kanamycin sensitive colonies all lacked amplification of the aph gene (data not shown).

Figure 3.

The pBS29 and pBS11 double transformed strain is unstable. Both pBS11 and pBS29 can excise from the mycobacterial chromosome in the presence of integrase.

Figure 4.

After excision of pBS11 and pBS29, pBS11 can reintegrate into the M. smegmatis chromosome blocking subsequent reintegration by pBS29 as the attB site is absent.

At the same time that bacteria were grown overnight in the absence of antibiotic, total DNA was prepared that may also carry small amounts of excised phage L5 integrative plasmid. E. coli were transformed with total DNA from M. smegmatis bearing pBS29 and pBS11 integrated vectors grown in the absence of antibiotic. If excised vectors are present within the total DNA they should transform E. coli and confer antibiotic resistance. A greater number of kanamycin resistant transformants (0.33/ng total DNA transformed) were recovered than hygromycin resistant transformants (0.17/ng total DNA transformed). Sixteen kanamycin resistant E. coli colonies containing plasmid were grown in liquid culture and plasmid DNA was purified using a miniprep procedure. The plasmids were then cut to verify similarity to the parent plasmids pBS29 and to reveal small plasmid differences with the frequent cutter Sau3AI. The results revealed identical cutting profiles for plasmids recovered from kanamycin resistant strains and pBS29 (Figure 5). Eight hygromycin resistant E. coli colonies were grown up and plasmid DNA was prepared. Plasmids were cut with Sau3AI and compared to cut pBS11. All plasmids prepared from the hygromycin resistant E. coli had an identical Sau3AI cutting profile as pBS11 (Figure 5). All plasmids recovered from kanamycin and hygromycin resistant E. coli could be retransformed into M. smegmatis to confer kanamycin and hygromycin resistance indicating that they were in fact pBS29 and pBS11 respectively. Thus it seems likely that integrated vectors are present within the mycobacterial cell as an extrachromosomal excised plasmid at least a portion of the time.

Figure 5.

The pBS29 and pBS11 as well as plasmids recovered from kanamycin and hygromycin resistant E. coli strains were cut with Sau3A1 and were visualized on a 2.0 % agarose gel.

Both pBS29 and pBS11 possess the integrase gene. Deletion of the integrase gene in pBS29 and pBS11 created the plasmids pBS33 and pBS37 respectively. M. smegmatis was transformed with pBS33 + pBluescriptint which has the integrase present on a mycobacterial non replicating vector. The pBS33 integrates in M. smegmatis bacterial cells and the nonreplicating integrase containing plasmid pBluescript is lost. The procedure is repeated with pBS37 and pBluescriptint. M. smegmatis with pBS33 and pBS37 integrated into the chromosome were grown for 20 hours in the absence of antibiotic. The mycobacteria were diluted and plated on 7H10 plates lacking antibiotic to obtain single colonies. Bacterial colonies were numbered and patched onto 7H10 + kanamycin or 7H10 + hygromycin plates. Removal of the integrase gene after M. smegmatis chromosomal integration by pBS33 or pBS37 completely prevented the loss of either plasmid as evidenced by the lack of development of kanamycin or hygromycin sensitivity during growth in the absence of antibiotic selection. One hundred individual colonies were tested. In addition, when total DNA was prepared from M. smegmatis bearing pBS33 and pBS37 and used to transform E. coli, this experiment resulted in no kanamycin resistant transformants/1.4 µg of total DNA transformed and no hygromycin resistant transformants/1.4 µg of total DNA transformed indicating that pBS33 and pBS37 are not present in an extrachromosomal form. Thus the process of in vivo excision and reintegration of phage L5 based integrating plasmids as well as the presence of extrachromosomal excised plasmid is dependent on the presence of the integrase gene product. In addition, it is extremely unlikely that the high rate of loss of kanamycin and hygromycin resistance in M. smegmatis with integrated pBS29 and pBS11 is due to mutation of the antibiotic resistance marker or some recombinitorial mechanism.

As stated above, kanamycin and hygromycin resistant plasmids were recovered from E. coli after transformation of total DNA prepared from M. smegmatis bacteria bearing integrated pBS29 and pBS11. One potential alternative explanation for the recovery of pBS29 and pBS11 plasmids is that pieces of chromosomal DNA containing the integrated vectors present in the total DNA were transformed into E. coli. Integrated plasmids present within chromosomal DNA could excise with the aid of de novo synthesis of the integrase protein within transformed E. coli. In this scenario one would expect each kanamycin resistant transformant to also be hygromycin resistant as excision of pBS29 requires either pre-excision of pBS11, or excision of pBS29 containing pBS11 integrated within this plasmid (Figure 1). In either case the E. coli strain would be both kanamycin and hygromycin resistant. Total DNA prepared from M. smegmatis containing integrated vectors pBS29 and pBS11 grown in the absence of antibiotic selection was used to transform E. coli. There were 52 kanamycin resistant transformants in E. coli patched onto hygromycin LB agar plates. Of the kanamycin resistant transformants none were also hygromycin resistant. Thus it seems likely that hygromycin resistant plasmids and kanamycin resistant plasmids are present within prepared total DNA as extrachromosomal structures that can be transformed separately into E. coli.

IV. Discussion

Phage L5 based integrating vectors are present within the bacterial cell in equilibrium between integrated and excised states. This is important because with selective pressure integrated plasmids can be lost from the bacterial cell in the presence of integrase. It had previously been shown that integrated plasmids are unstable within the host cell and this is probably due to integrase which was shown to function additionally as an excisionase (Lewis and Hatfull, 2000; Springer et al, 2001). Also, L5 integrated plasmids can be replaced by transformation of a mycobacterium containing a previously integrated vector with a novel second integrating vector (Pashley and Parish, 2003). It was determined that the original integrated vector is excised and then replaced with the new integrating vector at a fairly high rate (Pashley and Parish, 2003). This present study demonstrates that integrating vectors are found within mycobacteria in equilibrium between integrated and excised states by showing that pBS11 can excise and then reintegrate into the mycobacterial chromosome with some frequency.

Finally the fact that integrated vectors are present in equilibrium indicates that in a certain percentage of cells, excised vectors will be present. Thus investigators wanting to recover integrated vectors from mycobacteria can prepare total DNA and transform that DNA into competent E. coli cells to obtain the desired plasmids. In addition, if investigators wish to employ an L5 integrative vector with a deleted integrase and provide the integrase on a mycobacterial nonreplicating vector, the resultant integrated plasmid will be extremely stable. If investigators wish to recover the integrated plasmid, integrase can then be provided subsequently on a transformed replicating plasmid. Exposure of the integrated plasmid to the integrase supplied in trans should be sufficient to induce the integrating plasmid to again persist in equilibrium between integrated and excised states. Total DNA preparation should allow recovery of the integrated plasmid. Thus this will provide all the benefits of an extremely stable integrative plasmid while allowing for the ease of plasmid recovery.

Abbreviations

LB

Luria Broth

OD

optical density

PCR

polymerase chain reaction

Biography

Beatrice Saviola