|Home | About | Journals | Submit | Contact Us | Français|
The increasing interest in genetic manipulation of bacterial host metabolic pathways for protein or small molecule production has led to a need to add new genes to a chromosome quickly and easily without leaving behind a selectable marker. The present report describes a vector and four-day procedure that enable site-specific chromosomal insertion of cloned genes in a context insulated from external transcription, usable once in a construction series. The use of rhamnose-inducible transcription from rhaBp allows regulation of the inserted genes independently of the commonly used IPTG and arabinose strategies. Using lacZ as a reporter, we first show that expression from the rhamnose promoter is tightly regulatable, exhibiting very low leakage of background expression compared with background, and moderate rhamnose-induced expression compared with IPTG-induced expression from lacp. Second, the expression of a DNA methyltransferase was used to show that rhamnose regulation yielded on-off expression of this enzyme, such that a resident high-copy plasmid was either fully sensitive or fully resistant to isoschizomer restriction enzyme cleavage. In both cases, growth medium manipulation allows intermediate levels of expression. The vehicle can also be adapted as an ORF-cloning vector.
Genetic engineering of plasmids provides a rapid and flexible method of adding individual genes to a bacterial gene complement. However, multiple stable changes to the metabolic architecture of a strain are often desirable, for example providing new metabolic pathways, novel chaperone suites to allow proper folding of exotic target proteins, or exotic tRNA genes to allow smooth translation of foreign codons. Strain-building strategies thus quickly exhaust the repertoire of compatible replicons and drug-selection options. Direct manipulation of the genome itself then becomes an attractive strategy.
Genome manipulation strategies in bacteria at present range from whole-genome synthesis (1) to selective deletion or gene addition via homologous recombination using PCR products (2,3) and in favorable cases oligonucleotide mutagenesis in vivo (4). The first, while a technical tour-de-force, is not yet of routine utility. The second is very useful, especially due to generality, but a scar remains at the site of manipulation. The third is a scarless method useful for making small changes, but not for adding functions (4–6). A method less widely used (7) requires two cloning steps, additional steps for drug resistance removal and also leaves a scar. Older approaches that rely on homologous recombination are still in use but are more cumbersome (8).
Methods that focus on gene addition at defined neutral positions are useful complements to the replacement methods. These rely on site-specific recombinase functions. Lambdoid bacteriophages were the first site-specific systems used for adding genes to a bacterial chromosome (9–11). The utility of the site-specific approach has been expanded with a series of different phage integrase vectors, but each system leaves behind a drug-resistance marker, limiting its utility for strain building (12,13).
Transposon Tn7 has a site-specific mode of insertion, with a unique attachment site (attTn7) downstream of the highly conserved glmS gene. This system has the advantage that it inserts at very high frequency (14) into a single conserved site distributed through many taxa (15). A basic mini-Tn7 (mTn7) donor plasmid vector, pGRG36, was previously designed to deliver cargo to the specific site, attTn7 (16). The insertion (mobilization) procedure is straightforward: the desired DNA is placed in a multiple cloning site located between the Tn7R and Tn7L end sequences, at which the transposition proteins act to move the element. The donor thus constructed has (i) transposition genes separated from the region to be mobilized; (ii) a selectable marker for the donor plasmid, also separate; and (iii) a thermosensitive replicon, so that the donor is lost from the cell when grown at 42°C. When the donor is present in the cell at 30°C, chromosomal insertions accumulate due to the highly efficient action of the transposition proteins. Subsequent growth at 42°C and resulting plasmid curing allows recovery of strains carrying these chromosomal insertions at attTn7. In the model situation, 50–100% of colonies obtained at 42°C carry the mobilized element inserted at attTn7. It should be noted that the Tn7 system exhibits transposition immunity, so that insertion of a second element into an occupied site occurs at about a 60-fold lower frequency than the first insertion. With suitable transposase mutants the frequency of second insertions is only 3-fold lower than with wild type first insertions (17), although the stability of such tandem elements has not been investigated.
Some complications were anticipated in the use of the basic Tn7 donor plasmid pGRG36, stemming from expression signals in adjacent DNA both on the donor and in the chromosomal destination. Tn7R permits transcription to enter the element from the outside; insertion into chromosomal attTn7 disrupts the glmS transcription terminator, allowing readthrough into the element (17). This is an undesirable situation for the introduction of alien genes, particularly for those encoding toxic products. The expression context on the donor plasmid could also be problematic for the intermediate construct: adjacent to Tn7R is araC, reading into the element without a terminator; adjacent to Tn7L is ori-pSC101, with transcription also reading toward the element.
We based the design of the new vector, pMS26, on the backbone of pGRG36, with a new cloning region cassette between the sites of transposition protein action, Tn7L and Tn7R. To insulate expression of cloned genes from external signals, transcription and translation terminators are placed immediately inside the Tn7 ends. The design includes rhaBp, a rhamnose-regulatable promoter; convenient cloning sites and rapid protocol; and adaptability for use as a vector to clone open reading frames (ORFS) stripped of their native expression signals. ORF clones can then be expressed at the experimenter's direction using tightly regulated expression signals provided by the design for the cargo ORF. Due to the high efficiency and easy cloning strategy adopted, the present method has significant advantages in rapidity of construction. Performance was compared with familiar examples in E. coli using lacZ expression with or without its native promoter. In addition, complete plasmid protection via expression of the foreign methyltransferase, M.FnuDII (m5CGCG) was demonstrated using this cassette.
Genetic elements and fusions are designated according to the recommendations of ref. (18) and the American Society for Microbiology. For brevity, ‘mini-Tn7’ is abbreviated to ‘mTn7’; it signifies a genetic element mobilizable by the Tn7 transposition proteins. Mobilization results from action of those proteins on the two ends of the element and moves the element with all DNA between the ends to a new location. Each mTn7 (Table 1) is named here for the cargo that it carries, with an allele number distinguishing different variants of a similar construction. For example, mTn7Φ(rhaBp-lacZ)1 carries a fusion of the rhamnose promoter to the lacZ open reading frame (with one configuration of translation signal) while mTn7Φ(rhaBp-lacZ)2 carries a different such fusion (a different translation signal). mTn7Φ(rhaBp-lacZp-lacZ) carries the lacZ transcription unit downstream of the rhamnose promoter. USER: Uracil Specific Excision Reagent (NEB #M5505). USERBstBI: refers to the combination of Nt.BbvCI and BstBI that creates long extensions for efficient USER-enabled cloning described below. These extensions are not the same as those created by the combination of Nt.BbvCI and XbaI described for the original USER-enabled vector (19).
The elements listed in Table 1 were constructed by plasmid digestion and ligation with high concentration ligase (for synthetic segment received from Genscript), or processing of PCR products obtained using primers listed in Table 2 followed by USER assembly as described below. Junction sequences were verified for all constructions. Sequences were assembled and analyzed using the LaserGene suite from DNAStar. pMS26, pMS33 and pMS34 and sequences of all plasmids and mTn7 elements are available at Addgene (http://www.addgene.org/Elisabeth_Raleigh).
pGRG36 (16) was obtained from N. Craig (Johns Hopkins University School of Medicine). Oligonucleotides were obtained from IDT. The T7Et-rhaBp-rrnB1t cassette was obtained from Genscript. This cassette was cloned between the PacI and AscI sites of pGRG36 to create pMS26.
All molecular biology reagents [restriction enzymes, OneTaq Hot Start DNA polymerase, ligase (NEB202M), USER, and 2-log DNA ladder] were from New England Biolabs except Pfu Turbo Cx Hotstart DNA polymerase (Agilent Genomics).
Bacterial strains Derivatives of K-12 (NEB 10-beta and ER2683) and B (ER2566) are listed in Table 3.
Media were Luria-Bertani [Rich Broth (RB) (20)], LB (21) or M9 minimal salts (22). Minimal media were supplemented with glucose or rhamnose to 0.2%, or both glycerol and rhamnose to 0.2%. Drugs were ampicillin (100µg/ml) and chloramphenicol (5µg/ml for ER2566 derivatives, 15µg/ml for NEB 10-beta derivatives). Colonies were picked with capillary pipets into 0.85% NaCl for temporary storage before colony PCR or growth for testing of properties.
pMS26 was prepared for USERBstBI cloning according to the protocol below with the following additions and changes. Plasmid was purified with Qiaprep Spin miniprep kit (Qiagen); BstBI digestion was for 6h; the digested DNA was resuspended in 30µl 10mM Tris–HCl pH 8.5. The final concentration of linearized DNA was about 40ng/µl. The USER assembly reaction was carried out as described in Validation, below.
Complete digestion of the sites revealing long extensions is important. The processed vector can be stored at least for several months at −20°C. Any standard plasmid DNA purification can be used as long as a phenol/chloroform extraction and subsequent ethanol precipitation step are employed. We have found that Qiagen DNA column purifications may not be completely digested unless the DNA is phenol/chloroform extracted and ethanol precipitated. This protocol is for 10µg DNA but can, of course, be scaled up.
|10µg||pMS26 DNA (at least 640ng/µl)|
|Incubate 6h to overnight at 65°C.|
Fragments (100µl volume each PCR) to be cloned in pMS26 were prepared using 5U Pfu Turbo Cx Hotstart DNA polymerase, 0.25mM each dNTP, and the appropriate primers listed in Table 2 (0.5µM). Template for the PCR was either ~10ng MG1655 genomic DNA (for lacZ constructs) or ~1ng Qiagen plasmid miniprep DNA from ER3169 (for fnuDIIM chromosomal insertions). Plasmid PCR template was first linearized with EcoRI to minimize plasmid carryover during subsequent USER cloning. PCR cycling parameters were as recommended by Agilent. Following PCR, samples were purified using the QIAquick PCR purification kit (Qiagen) prior to the USER assembly reaction.
An expedited protocol was developed that can be used to rapidly isolate multiple independent constructs and chromosomal insertions from them. Recovery of chromosomal insertions is lower (~10%) with this element and protocol than the original procedure (50–70%), but faster (4 days from PCR to identified insertion) and more automatable.
Colony PCR: reactions for colony PCR screening of the 42°C colonies for chromosomal insertions contained 2.5U OneTaq Hot Start DNA polymerase (NEB) in 50µl with primers 7 and 8 (Table 2) (0.5µM each). 100µl aliquots of 0.85% saline colony suspensions were heated to 99°C for 5min to prepare extracts; 5µl of each extract was used for PCR. PCR cycling parameters were as recommended by NEB.
Assays for β-galactosidase were a modification of the Miller protocol (22). Briefly, cultures were grown from three single colonies per strain in the medium specified; A600 was assessed directly on mid-log cultures or on a 1:5 or 1:10 dilution of late-log or stationary phase cultures to obtain readings between 0.2 and 0.8; samples were collected by centrifugation and resuspended in the same volume of lysis buffer (recipe); debris was clarified by centrifugation, and 1–50µl of the extract was used in 1ml reactions with ortho-nitrophenyl-β-d-galactoside substrate as described in (22). Miller units were calculated as 1000×A420/(volume×time×A600).
Restriction digests for determination of M.FnuDII activity: ER3239 (negative control) and ER3240 were tested for the ability to modify DNA in vivo using as reporter a co-resident high-copy plasmid with an easily diagnosed banding pattern. A total of 10µl of miniprep DNA prepared from cells grown under various induction regimens was digested with 3U of AclI in 1×NEBuffer 4 supplemented with 100µg/ml BSA at 37°C for 90min to linearize the plasmid at a convenient location; 10U of BstUI was added and incubation continued at 60°C for one h to produce fragments diagnostic of unmodified sites.
The vector pMS26 receives a desired gene within a mobile element (Figure 1, Steps 1–3) that inserts with high efficiency at a unique site in the genome of E. coli (Figure 1, middle). The donor thus assembled is introduced into cells by transformation (Step 4) and selected at low temperature for ampicillin resistance encoded by the vector. Transformants are purified at high temperature (Step 5), leading to loss of the temperature-sensitive replicon along with drug resistance and transposition protein genes. 10% of the survivors carry the mobile element downstream of glmS (Figure 1, bottom), disrupting its terminator, glmSt (Figure 2). The cloned segment is protected from expression driven by external signals with transcription and translation terminators flanking the cloning region (Figure 3A). Inducible expression is provided by the rhamnose-regulated promoter rhaBp (Figure 3A) characterized by very tight control. Uninduced expression is undetectable; maximal expression [activated by RhaS and CRP (23)] is good, around 10% of fully induced lacp. A cloning strategy that relies on long single-stranded extensions (Figure 1, Step 1, USERBstBI; ‘Materials and Methods’ section) allows this to be accomplished in four days, without isolation of the vector intermediate to verify its structure. The orientation of expression has been designed to minimize potential conflict with adjacent essential gene glmS (Figure 2), in case transcription emanating from the element does not terminate completely. Restriction enzyme sites within the element enable construction of alternative expression environments as desired (Figure 3B) while retaining protection from external transcription and translation.
A preliminary trial using pGRG36 revealed problems with transcription entering the mobile element from vector or chromosome (data not shown). Transcription from glmS on the chromosome and from araC on the vector will read into the element from the right (Figure 2 showing the E. coli chromosomal region; vector not shown). Transcription from repA on the vector will read in from the left (not shown); but ptsS reads away on the chromosome (Figure 2). Accordingly, design of pMS26 provides insulation from outside promoters and from translational readthrough both in the plasmid vector and in the chromosomal location. This is accomplished with three-frame stop codons inside each border, and transcription terminators (rrnBt inside Tn7R, the T7 early terminator T7Et inside Tn7L) as displayed in Figure 3A. The T7Et terminator is NOT adapted to terminate T7 RNAP; rather it functions in termination of early transcripts read by the E. coli RNAP (24). Here, it is oriented to terminate incoming transcription. The rrnBt terminator is reported to act bidirectionally (25). Providing regulated expression with a very low background was an additional goal; the sequence of the rhaB regulatory region was designed based on reference (23) and corresponds to bases −289 to −99 relative to the transcription start of rhaBAD in Fig. 1 of that reference. Use of cloning and expression options provided is discussed further below.
Efficient USER-facilitated cloning can be used to clone downstream of rhaBp. See ‘Materials and Methods’ section for a description of how to process the vector for this purpose (Step 1. USERBstBI Vector Preparation).
Cloning with 8–12 basepair single-stranded extensions substantially improves the efficient recovery of inserts (19). Long single-strand extensions on insert DNA fragments can be created by employing uracil-containing PCR primers, followed by excision of uracil, and cleavage at the abasic site thus created. The excision and cleavage reactions are accomplished with the USER enzyme mixture (19). In the original report, ≥95% of transformants contain the designed clone; this efficiency has been reproduced here.
USERBstBI extensions: Creation of long extensions on the vector requires use of suitably disposed sites for a type II restriction enzyme and a nicking enzyme; in the original configuration, this was achieved with XbaI and Nt.BbvCI, each of which must be unique in the cloning vector. In the present instance, the starting vector pGRG36 had no BbvCI sites, but numerous XbaI sites were present. The type II enzyme must have a T 5′ to its cleavage position for the strategy to work; we identified BstBI (TT|CGAA) as a substitute for XbaI (T|CTAGA) that would be unique in the vector. For convenience, we call the resulting extensions USERBstBI extensions. Since BstBI cleaves after the second base of its site instead of after the first, use of this enzyme for vector preparation instead of XbaI necessitates adding an extra T before the target U to the PCR primer when amplifying the desired insert (see below).
To use this vector processed as described in ‘Materials and Methods’ section, design PCR primers for your favorite gene according to the example in Figure 4. (Note that these primers cannot be used with pNEB206A due to the extra base.)
Upstream forward primer (N-terminal side, for transcriptional fusion):
5′GGAGACATU-(optionally translation signals)-your-upstream-primer-binding-site3′
Downstream reverse complement (C-terminal side):
Transcription unit cloning: If the goal of the experiment is to use the native promoter and regulatory sites for expression, the primer binding sites should be upstream of that transcription unit. UserBstBI cloning will result in rhaBp upstream in tandem with the native promoter. In the absence of rhamnose, rhaBp is essentially silent.
Translation unit cloning: If rhaBp-directed expression is desired, the primer-binding sites should be upstream of translation signals (essentially the 5′ untranslated region or 5′-UTR) for your gene, or else translation signals should be designed into the primers. Two options for translation signals are described in the results below and displayed in Figure 4.
Efficient insertion results from PCR with primers described above and in Figure 4. These contain strategically located uracil, which can be acted on by the USER reagent to reveal 3′ single-strand extensions compatible with the processed vector (step 3 in Figure 1). For complex assemblies of multiple fragments, ligation and gel purification is recommended (19). However, if a single amplicon contains all the genes of interest, ligation is not needed and the assembly mixture can be transformed directly, plated on ampicillin and incubated at low temperature to allow vector replication. Expression of transposition proteins during colony growth results in mobilization of the mTn7 to its chromosomal target, attTn7 (Figure 1, middle). Repurification of transformant colonies at 42°C in the absence of ampicillin (Step 5 of Figure 1) results in efficient plasmid loss. The survivors all lack the plasmid, and about 10% carry mTn7 inserted at attTn7.
Alternatives to the USER cloning protocol and rhaBp expression are also enabled in pMS26. At least three expression configurations can be achieved using restriction enzyme sites in the vectors described here:
Alternatively, plasmids that already have translation signals (described below, pMS33 or pMS34) can be adapted as ORF-cloning vectors with lac stuffer fragment, allowing rhaBp-directed expression. The lacZ fragment can be replaced with an alternative by digesting with AatII at the N-terminal end of the coding sequence and PmeI, NotI or AscI at the C-terminal end. Blue-white screening with Xgal in the presence of rhamnose should identify replacement clones.
In constructing ER3228 containing mTn7Φ(rhaBp-lacZp-lacZ)1, 0.06 pmole PCR product obtained using primers 1 and 2 of Table 2 (10µl of 0.006ng/µl) and 0.005pmole (1µl, 40ng) processed pMS26 were incubated with 1 unit of USER enzyme in a total volume of 12µl for 15min at 37°C, and then incubated 15min at room temperature. 3µl of the assembly reaction (10ng of pMS26) was mixed immediately with 50µl commercial competent ER3019 host cells, incubated on ice for 30min, at 37°C for 2min, and finally on ice for 5min. Cells were gently mixed with 1ml LB and outgrown at 30°C for 1h. 100µl of undiluted, 10-fold diluted and 100-fold diluted culture was plated on LB with 100µg/ml ampicillin and incubated 24h at 30°C. 7300 transformants/ml were obtained (7×105 perµg processed vector). When the transformation mix was replated at 30°C with IPTG-Xgal present, ~3/5000 colonies were white, indicating a very low frequency of vector without insert.
With lesser insert-vector ratios and cells of lower competence, 20–200 colonies typically resulted from plating 10% of the transformation mix (~103–04 transformants perµg processed vector).
Table 4 shows five trials in two strains employing the rapid insertion protocol. On the day following transformation of each USER assembly reaction, 7–20 transformants were streaked on to LB plates without drug and incubated overnight at 42°C to cure the plasmid.
Colonies picked from the 42°C plates were analyzed by colony PCR. Primers 7 and 8 (Table 2) yielded a product of 841 bp for an empty attTn7 site, while colonies with lacZ or fnuDIIM inserts display products of 5129 and 3201 bp, respectively. The observed ratio # inserts/# tested colonies are given in Table 4, lines 1–5. Trials with two configurations of lacZ and one of fnuDIIM as cargo using the pMS26 vector in two hosts consistently yielded an insertion frequency of ~10% (Table 4).
We considered the possibility that additional generations of growth in the presence of the donor, or simply longer time, might yield a higher frequency of chromosomal insertion. To test this, in one experiment transformants (~107cells/colony) were patched onto LB Amp, incubated overnight at 30°C, then a plug (~107 cells/plug) from the patch was grown in 1ml RB at 30°C overnight (~109total cells) then diluted and plated at 42°C overnight to cure the plasmid. This longer protocol had no effect on the fraction of descendants with insertions (Line 6 of Table 4).
When the entire lacZ expression region is included in the cloned fragment, an appropriate 100-fold IPTG-dependent induction ratio is seen (compare lines 4 and 6 of Table 5). Rhamnose addition did not significantly increase β-galactosidase expression above the background from uninduced lacp (compare line 4, Table 5 with line 5, Table 5), suggesting that LacI binding may block RNAP transcription from rhaBp in this configuration. The nearest LacI binding site is 31bp from the rhaBp transcription initiation site. The background (level in the absence of any element, 100–200U, lines 1–3) is indistinguishable from zero; note that the detection limit can be reduced by using more cells and longer incubation times (increasing the denominator in calculating Miller units).
Expression of β-galactosidase in the mTn7(MCS MS26) context was evaluated in two experiments using rhaBp, USER cloning and the two different 5′-UTRs illustrated in Figure 4, with results shown in Figure 5. The longer version [mTn7Φ(rhaBp-lacZ)1; LTS] of the 5′-UTR consistently yielded 2- to 3-fold more activity than the short version [mTn7Φ(rhaBp-lacZ)2; STS]. Expression driven by rhaBp in this context (in minimal medium) is maximally about 10-fold lower than fully induced lacZ (in RB; compare with Table 5).
Choice of growth media allows modulation of expression by rhamnose over a 1000-fold range, from the combined effect of catabolite activation and rhamnose activation. Table 6 shows that uninduced and glucose-repressed expression using STS (lines 7–9) is indistinguishable from the background with no lacZ (lines 1–3), while LTS may be slightly higher (line 4–6). Rhamnose results (lines 10–15) are the same data as experiment 2 of Figure 4, showing improved expression mediated by the LTS. Growth supported by glycerol together with rhamnose yields somewhat lower specific activity than rhamnose alone (line 16–18 of Table 6; Miller units are normalized for A600, a proxy for protein content). This mixture of carbon sources both subject to catabolite activation, in which both the rhamnose and glycerol regulons are sharing the positive activator CRP, may lead to lower expression of each regulon due to competition for activator. Despite the lower specific activity, the growth properties of all tested strains are considerably better when glycerol is present with or without rhamnose (shorter lag, faster growth and higher density; growth curves not shown).
Regulated expression is not limited to the reporter lacZ that was used for literature comparison. A strain constructed to drive expression of the FnuDII DNA methyltransferase from the rhaB promoter also showed large sugar-dependent expression changes. The substrate pMC63 is a high-copy plasmid with two CGCG sites, which can be modified by M.FnuDII or cleaved by BstUI (26). Plasmid-encoded CGCG sites modified by the methyltransferase (during cellular propagation in this experiment) are protected from BstUI cleavage (27). Figure 6 illustrates the protective action in vivo, mediated by induction of the methyltransferase in ER3240, containing pMC63 as a reporter and mTn7-Φ(fnuDIIM)1 as an expression construct. When this plasmid was grown in the presence of fnuDIIM in M9 rhamnose, the in vivo expression levels were sufficient to fully modify these CGCG sites [no cleavage of the AclI-linearized extracted plasmid, lanes (minimal, R, +)]. On the other hand, the in vivo expression levels only partially modified the CGCG sites within pMC63 when grown in rich medium with rhamnose (rich, R, +). Glucose-grown cultures did not protect at all (minimal, G, +; rich G, +): the digest patterns match those from cultures entirely lacking the mTn7Φ(fnuDIIM)1 element (minimal, G, −; minimal, R, −; rich, G, −, rich, R, −).
The complex transposon Tn7 has several features that can be adapted for genetic tool development. In the present study, the high-specificity, high-efficiency pathway targeting attTn7 is the aspect of interest. This pathway, directed by the target selector protein TnsD together with the TnsABC catalytic assembly, mediates insertion into one (E. coli, Yersinia, Acinetobacter, Francisella) or a small number (Burkholderia) of sites in a broad array of bacteria (28). Even human DNA has a high-efficiency site, also near a gene encoding N-acetyl glucosamine synthase (29). This highly conserved gene (glmS in bacteria) provides the widely distributed recognition site, but since the site of actual insertion is outside the coding region, downstream of the recognition site, insertion is well-tolerated.
The arrangement presented here mitigates some of the problems associated with using the original pGRG36 system. Genes within the mobile element are insulated from expression signals originating outside of it, and the high-efficiency cloning approach reduces problems associated with manipulation of large plasmids.
High efficiency insertion allows screening (examination one-by-one) rather than selection (using drug selection to eliminate unproductive background) for the construct desired. The mobile element mTn7(MCS MS26), carried on pMS26, still retains an insertion efficiency high enough to allow screening rather than selection for chromosomal insertions (10% of 42°C survivors; Table 4). Although ampicillin is used to select for the presence of the initial construct, that construct is efficiently lost at 42°C, and drug-sensitivity is restored in the final strain. The streamlined insertion protocol was effective in two host strains of very different pedigrees, with three constructs of distinct function. Possibly the insertion efficiency might be further increased by adding arabinose to the ampicillin plates used to select transformants (Step 4 of Figure 1).
The mobile element mTn7(MCS MS26) presumably could also be mobilized via the TnsE-dependent random insertion pathway that targets conjugal plasmids. We have not tested this lower efficiency transposition pathway. TnsE is not encoded on pMS26 but could be provided in trans.
We compared the rhaBp expression system to the well-characterized lac expression system in several ways. One trial (Table 5) examined lacI-regulated expression of the lacZ transcription unit, in the context of the mTn7(MCS MS26) element. In this situation lacZp directed transcription and was regulated by LacI encoded on the ϕ80 Δ(lacZ)M15 prophage resident in the host. The mTn7Φ(rhaBp-lacZp-lacZ)1 element contains, downstream of rhaBp, a segment of wild-type lac operon sequence extending from the end of lacI to the beginning of lacY, including all annotated regulatory elements. In the absence of rhamnose, IPTG mediates a 100-fold induction, similar to induction seen in the wild-type lac operon (not counting glucose repression, not measured here) and in other lacZ reporter controls [see, e.g. (30); lacUV5p was examined in that case].
Also notable was the lack of rhamnose induction above the background expression from uninduced lacp. The rhaBp transcription start is 31 bp from the closest (furthest upstream of lacp) LacI binding site. Several mechanisms by which bound regulators could interfere with each other or with RNAP can be imagined.
The provision of a tightly controlled inducible rhaBp promoter together with a strong translation initiation region adds flexibility. Expression is very low in the presence of glucose [Table 6, lines 4–9 and Figure 6 lanes (G, +)], indistinguishable from the negative control (Table 6, lines 1–3; Figure 6, lanes (G, –; R, –)]. The highest expression was achieved by combining the strong translation signal (LTS) in M9 medium with rhamnose as sole carbon source (Table 6, lines 10–12; Figure 5, LTS; Figure 6, M9, R, +). The level of β-galactosidase activity was ~10% of fully induced lacZ from its own promoter (compare Table 5, IPTG with Table 6, lines 10–12). For reasons we have not explored, growth under these conditions is slow, and final cell yields are half that with glucose as carbon source. Induction of the rhamnose operon is known to be slow (31), possibly accounting for the long growth lag.
The donor plasmid pMS26 also was designed to accommodate alternative cloning and expression strategies, as described in the section on vector design. Such modified protocols will still enjoy the insulation from external expression and the high insertion specificity of the mobile element described here.
Funding for open access charge: New England Biolabs.
Conflict of interest statement. None declared.
We thank Nancy Craig for sharing pGRG36 and advice related to Tn7 behavior. We are grateful for comments and advice from Mehmet Berkmen, Yuri Londer, Jim Samuelson, Rich Roberts, Carine Robichon, Peter Weigele and Yu Zheng and constructive criticism from anonymous reviewers. We are especially grateful to Don Comb for underwriting this research and the many years of projects from which it stemmed.