OUP user menu

Use of stabilized luciferase-expressing plasmids to examine in vivo-induced promoters in the Vibrio cholerae vaccine strain CVD 103-HgR

Cara E. Morin, James B. Kaper
DOI: http://dx.doi.org/10.1111/j.1574-695X.2009.00580.x 69-79 First published online: 1 October 2009


Live, attenuated Vibrio cholerae vaccines can induce potent immune responses after only a single oral dose. The strategy of harnessing these strains to present antigens from heterologous pathogens to the mucosal immune system shows great promise. To fully realize this possibility, V. cholerae strains must be created that stably express antigens in vivo in sufficient quantity to generate an immune response. In vivo-induced promoters have been shown to increase the stability and immunogenicity of foreign antigens expressed from multicopy plasmids. We report the construction of a series of genetically stabilized plasmids expressing luciferase as a heterologous protein from the following in vivo-induced promoters: V. cholerae PargC, PfhuC and Pvca1008, and Salmonella enterica serovar Typhi PompC. We demonstrate that several of these expression plasmids meet two critical criteria for V. cholerae live vector vaccine studies. First, the plasmids are highly stable in the V. cholerae vaccine strain CVD 103-HgR at low copy number, in the absence of selective pressure. Second, real-time bioluminescent imaging (BLI) demonstrates inducible in vivo expression of the promoters in the suckling mouse model of V. cholerae colonization. Moreover, the use of BLI allows for direct quantitative comparison of in vivo expression from four different promoters at various time points.

  • Vibrio cholerae
  • bioluminescent imaging
  • promoter
  • vector vaccine
  • in vivo expression


Live attenuated bacterial vaccines offer a promising opportunity to deliver heterologous antigens to the immunological inductive sites involved in a wild-type human infection. Several species of enteric bacteria including Salmonella enterica serovar Typhi and Shigella flexneri have been used to express antigens from a wide variety of unrelated pathogens. Delivering antigens in this way induces immune responses against both the bacterial host strain and the foreign antigen(s), and has the advantage of oral delivery, increasing safety and compliance.

Vibrio cholerae offers several potential advantages as a live vector vaccine: (1) V. cholerae is noninvasive, but induces excellent mucosal and systemic immune responses often after just one dose of live vaccine; (2) most attenuated strains of V. cholerae maintain the highly immunogenic B subunit of the primary enterotoxin, cholera toxin, which has been demonstrated to be an excellent adjuvant for heterologous antigens (Holmgren et al., 2005); (3) attenuated strains of V. cholerae expressing and secreting several antigens can elicit heterologous immunity that is protective in animal models (reviewed in Silva et al., 2008); and (4) several live attenuated strains have been demonstrated to be safe and immunogenic in human subjects (reviewed in Tacket & Sack, 2008; Holmgren & Kaper, 2009), including one licensed for human use, CVD 103-HgR. The safety and immunogenicity of CVD 103-HgR has been established in a number of randomized, placebo-controlled, double-blind clinical trials throughout the world, which makes it an ideal vaccine platform for expression and delivery of heterologous antigens (Cryz et al., 1990, 1992, 1995; Kotloff et al., 1992; Tacket et al., 1992, 1999).

A critical balance must be maintained in a live vector vaccine whereby sufficient levels of foreign antigen are expressed to elicit an immune response without overattenuating the bacterial host strain (Bumann, 2001; Galen & Levine, 2001). Two major factors influencing expression of genes encoding heterologous antigens are the plasmid copy number and the choice of the promoter. Foreign genes expressed from multicopy plasmids may be unstable in vivo. One way to enhance the stability of plasmid-based heterologous expression systems is to use low-copy-number plasmids. However, this raises concerns about producing enough heterologous protein in vivo to mount an immune response.

An additional factor in plasmid stability is the choice of promoter for heterologous protein expression. Several promoters have been used previously to express heterologous antigens in V. cholerae vaccine strains, including constitutively active tac, lac and trc promoters. However, constitutive expression of high levels of foreign antigen may be toxic to the live vector. The use of in vivo-inducible (ivi) promoters provides a possible solution to the problem of toxicity or metabolic burden of the heterologous antigen on the host bacterial strain. ivi promoters that exhibit low expression in vitro allow for optimal growth conditions for preparation of inocula, while maximizing expression in vivo. Early examples of V. cholerae promoters used to express foreign antigens include heat-shock-regulated htpG (Butterton et al., 1997; John et al., 2000), iron-regulated irgA (John et al., 2000), anaerobically induced nirB (Chen et al., 1998; Fontana et al., 2000) and the cholera toxin promoter (Fontana et al., 2000; Liang et al., 2003). Several of these strains elicited immune responses against the heterologous antigen, but plasmid instability in vivo was suspected to reduce immunity.

Several genomic-level screens have identified new genes that are activated in vivo after V. cholerae infection of animals and humans (Chiang & Mekalanos, 1998; Lee et al., 2001; Merrell et al., 2002a, b; Hang et al., 2003; Larocque et al., 2005, 2008; Osorio et al., 2005). One recent in vivo expression technology (IVET) study conducted in humans led to the identification of several previously unreported and highly inducible ivi promoters with great potential for use in a foreign antigen expression system (Lombardo et al., 2007). Here, we exploit several of these novel ivi promoters for expression of bioluminescence using genetically stabilized expression plasmids, originally developed for use in attenuated S. Typhi and S. flexneri live vectors (Galen et al., 1997, 1999; Altboum et al., 2001, 2003; Stokes et al., 2007). Bioluminescent imaging (BLI) is a highly sensitive technique used previously in mice to follow colonization dynamics noninvasively, leading to the identification of previously unknown sites of pathogen colonization (Hardy et al., 2004; Wiles et al., 2006). Here, we use this powerful imaging technique to quantitatively compare the in vivo induction of several V. cholerae ivi promoters previously identified by IVET to be induced in human volunteers.

Materials and methods

Bacterial strains and growth conditions

CVD 103-HgR (O1, classical biotype) was used for all expression studies. N16961 (O1, El Tor biotype) was used as a template for PCR amplification of promoter sequences. Cells were grown in 1 × M9 salts plus 20% glucose as a carbon source, Luria–Bertani (LB), 3-(N-morpholino)propane sulfonate (MOPS) or EZ-Rich (Teknova). Kanamycin was used at a concentration of 50 µg mL−1 to maintain plasmids in vitro, except where noted.

Construction of luciferase expression plasmids

Construction of pOMP15lux, pOMP5lux and pOMP60lux

All plasmids are listed in Table 1. Because our intention was to insert several different promoters upstream of the lux operon, our first goal was to create an easily replaceable cassette to move promoters into and out of this plasmid. To this end, pGEN-luxCDABE (a generous gift from M. Lane and H. Mobley) was digested with EcoRI and SnaBI to remove the EM7 promoter, which was replaced by PompC from pSEC84. To accomplish this, PompC was removed from pSEC84 by digestion with SpeI, end filled with T4 DNA polymerase and subsequently digested with EcoRI. This PompC fragment was then ligated into pGEN-luxCDABE to create pOMP15lux. We then subcloned the entire PompC-lux fragment from pOMP15lux as an c. 6500-bp EcoRI–NheI cassette into appropriately digested pSEC10 (ori101) and pSEC84 (oriE1) to create pOMP5lux and pOMP60lux, respectively.

View this table:
Table 1

Plasmids used in this study with relevant features noted

PlasmidApproximate copy no.PromoterRelevant characteristicsSources or references
pGENlux15Pem7ori15A luxCDABE bla hok-sok par parALane et al. (2007)
pSEC105PompCori101 clyA aph hok-sok par parAStokes et al. (2007)
pSEC8460PompCoriE1 clyA aph hok-sok par parAGalen et al. (2004)
pOMP5lux5PompCori101 luxCDABE aph hok-sok par parAThis work
pOMP15lux15PompCori15A luxCDABE bla hok-sok par parAThis work
pOMP60lux60PompCoriE1 luxCDABE aph hok-sok par parAThis work
pCM105Pnegori101 luxCDABE aph hok-sok par parA rrnBThis work
pCM115PargCori101 luxCDABE aph hok-sok par parA rrnBThis work
pCM135PfhuCori101 luxCDABE aph hok-sok par parA rrnBThis work
pCM145Pvca1008ori101 luxCDABE aph hok-sok par parA rrnBThis work
pCM175PompCori101 luxCDABE aph hok-sok par parA rrnBThis work
pCM1860PnegoriE1 luxCDABE aph hok-sok par parA rrnBThis work
pCM1960PargCoriE1 luxCDABE aph hok-sok par parA rrnBThis work
pCM2260Pvca1008oriE1 luxCDABE aph hok-sok par parA rrnBThis work

Construction of pCM10

To reduce background luminescence resulting from read-through transcription into the lux operon, a copy of the rrnB terminator was PCR amplified from pSE380 (Invitrogen) using primers K5917 and K5918 (Table 2) to construct a cassette flanked at the 5′-terminus by MfeI and at the 3′-terminus with EcoRI and BamHI sites. The resulting PCR product was ligated into pOMP5lux digested with EcoRI and BamHI, creating the five-copy pCM10 promoterless negative control plasmid, which was confirmed to exhibit extremely low background luminescence.

View this table:
Table 2

Primers used in this study

Primer no.TemplateCassette createdDirectionOligonucleotide sequence (5′→3′)
K4125N16961 genomic DNAPargCForwardcaggatcctaagtggtgaaagagagcgaaca
K4126N16961 genomic DNAPargCReversetgggaattcgcggccgctatctttgctcctgctattgacagagtg
K4245N16961 genomic DNAPvca1008Forwardgaattcgcggccgctaatttcgattcctcatcactcaccacggcg
K4246N16961 genomic DNAPvca1008Reverseggatcctattagtttgattgggctttac
K5512N16961 genomic DNAPfhuCForwardgaattcgcggccgcggtgccttaatgaaagtaattg
K5513N16961 genomic DNAPfhuCReverseggatccattgcgcatgataaaacctctgc
K5917pSE380 (Invitrogen)rrnB-PnegForwardcaattgataaaacagaatttgcctgg
K5918pSE380 (Invitrogen)rrnB-PnegReverseggatccattcttatgaattcgagtttgtagaaacgcaaaaag
  • Restriction enzymes sites are underlined.

Construction of five- and 60-copy expression plasmids

PargC, PfhuC and Pvca1008 promoters were PCR amplified from wild-type V. cholerae N16961 genomic DNA to create cassettes with 5′-terminal EcoRI–NotI sites and BamHI at the 3′ terminus. Primers used are listed in Table 2. Promoter fragments were inserted as EcoRI–BamHI cassettes into pCM10 digested with the same enzymes, thereby creating five-copy plasmids with the argC (pCM11), fhuC (pCM13) and vca1008 (pCM14) promoters. The ompC promoter was also removed from pSEC84 as an EcoRI–BamHI cassette and inserted into appropriately digested pCM10 to create the five-copy pCM17. The terminator–promoter cassette from pCM10, pCM11 and pCM14 was then removed by digestion with XhoI and BamHI and ligated into pOMP60lux digested with the same enzymes to create the 60-copy isogenic plasmids pCM18, pCM19 and pCM22, respectively. The expected nucleotide sequence of all expression plasmids was verified by sequence analysis.

In vitro induction studies

Expression plasmids were introduced into V. cholerae by electroporation as described previously (Marcus et al., 1990). For experiments reported in Fig. 1, plasmid-bearing V. cholerae live vectors were subcultured 1 : 500 from overnight cultures (using two to three colonies from freshly streaked plates) into fresh medium and added in triplicate (200 µL per well) to an optical 96-well plate. Absorbance and luminescence of 200-µL cultures grown at 37 °C in an optical 96-well plate were measured at 20-min intervals using a Synergy HT luminometer. Triplicate or quadruplicate samples were assayed for each strain and medium condition.

Figure 1

Luciferase kinetics and stability. The symbols shown in (a) are used for both panels. The symbols for five-copy plasmids are solid and those for 60-copy plasmids are open. (a) and (b) show strains grown in 96-well plates in minimal and rich media, respectively. Each value is the mean of triplicate cultures and is representative of at least three experiments. The values of 60-copy PargC strain are cut-off at 500 min; this strain goes on to achieve a peak level of 3.94 × 106 at 780 min and then rapidly declines like the other strains.

For experiments reported in Fig. 2, overnight cultures grown from two to three freshly isolated colonies were diluted 1 : 100 into fresh LB and grown to midlog phase (OD600 nm∼0.4). Several 1-mL portions of each culture were pelleted and resuspended in an equal volume of minimal medium supplemented with 0.5 mM arginine, 50 µM 2,2′-dipyridyl, or 50 µM FeSO4 and added in triplicate (200 µL per well) to an optical 96-well plate. Bioluminescence and absorbance were again measured every 20 min using a Synergy HT luminometer.

Figure 2

In vitro regulation of promoters. Strains were grown in LB to mid-log phase and then resuspended in permissive or nonpermissive medium. (a) and (b) show the relief of repression of the argC promoter by the removal of exogenous arginine in the (a) five- or (b) 60-copy plasmids. (c) Repression of the fhuC promoter by the addition of FeSO4. Each value is the mean of triplicate cultures; error bars show SD. Each graph is representative of at least three experiments.

In vitro stability studies

In vitro stability experiments were performed essentially as described by Fang (2008). Briefly, two to three freshly isolated colonies were used to establish starter cultures grown at 37 °C in LB without selection. Twenty-four hours later, the cultures were diluted and plated onto nonselective medium (day 0) and reinoculated into fresh LB at a 1 : 100 dilution. This process was repeated each day for up to 5 days. Plate images were acquired using an IVIS 200 imaging system at small binning for 10–30 s to determine the percentage of luminescent colonies.

In vivo BLI

Freshly isolated colonies were resuspended in 10 mL of phosphate-buffered saline (PBS) and adjusted to an OD600 nm of 0.2 (c. 1 × 108 CFU mL−1). Evans blue dye (40 µL) was then added to 1 mL of each culture. Four- to five-day-old CD-1 mice were isolated 1 h before inoculations. Mice were individually anesthetized using 3% isoflurane, intragastrically inoculated with 50 µL of culture through flexible silicon tubing and imaged immediately for 1 min with medium binning in an IVIS 200 imaging system. Luminescence within the mice was calculated by generating an identically sized region of interest circle for all mice using living image 3.0 software. Total photon flux (photons s−1) was used for all calculations. At various time points, the mice were again anesthetized and imaged individually with the same parameters. At the final time point, mice were euthanized immediately after imaging by cervical dislocation and their intestines were dissected and submerged in 2 mL PBS+10% glycerol. After homogenization, the intestinal samples were diluted and plated to determine CFU. Colony plates were imaged as described above to confirm in vivo stability of plasmids. This animal protocol was approved by the Institutional Animal Care and Use Committee at the University of Maryland School of Medicine.

Results and discussion

Construction of plasmids expressing luciferase under ivi promoter control

To assess the expression of ivi promoters in vivo, we elected to use lux-mediated bioluminescence as a reporter. Bioluminescence is a useful and quantitative tool for studying gene expression of pathogens in vivo in animal models of infection (reviewed in Contag, 2008). Foreign antigens delivered by live vectors are usually expressed using multicopy plasmids such that high levels of antigen are presented to the immune system. However, such plasmids are often unstable and we therefore elected to express the lux operon from genetically stabilized expression plasmids. These plasmids encode a maintenance system comprising the Hok-Sok postsegregational killing function and two plasmid-partitioning loci to enhance plasmid inheritance (Galen et al., 1999; Lane et al., 2007). The plasmid pGEN-luxCDABE (Pem7, c. 15 copies per cell; Lane et al., 2007) was re-engineered for this study to create a removable promoter cassette and a copy of the rrnB terminator was inserted upstream of the promoter cassette to eliminate background luminescence observed in the absence of the terminator (data not shown). This re-engineered terminator-EcoRI-Pneg-BamHI-lux operon cassette was then introduced into low-copy (c. 5 copies per cell) and high-copy (c. 60 copies per cell) genetically stabilized plasmids creating promoterless negative controls (Pneg5 and Pneg60).

We then replaced Pneg with one of three ivi promoters identified by IVET in human volunteers (Lombardo et al., 2007). Two of the promoters, PargC and Pvca1008, had also been previously identified in a suckling mouse IVET study (Osorio et al., 2005; Lombardo et al., 2007). These two promoters are induced in the suckling mouse model at levels of 700- and 770-fold, respectively, compared with growth in vitro in rich broth medium. We chose a third promoter from the human IVET study, PfhuC, because a strain with a mutation in this gene was shown in follow-up experiments to display a colonization defect in the suckling mouse intestine. A fourth promoter, PompC, from Escherichia coli was also included in our studies. OmpC is a porin that is present in many gram-negative bacteria, but not in V. cholerae. In E. coli, ompC expression is induced by high osmolarity, and it encodes one of the most highly expressed outer membrane proteins. PompC has been used successfully in both attenuated S. Typhi and Shigella strains to express heterologous antigens and has been shown to induce immune responses in animal models (Altboum et al., 2001, 2003; Galen et al., 2004; Stokes et al., 2007).

All plasmids were electroporated into a spontaneous streptomycin-resistant derivative of the attenuated V. cholerae strain CVD 103-HgR. The safety and immunogenicity of CVD 103-HgR has been tested in thousands of individuals, including children and persons with HIV, making it an ideal V. cholerae vector vaccine (reviewed in Tacket & Sack, 2008).

Evaluating in vitro expression

A major goal in live vector vaccine construction is to limit expression of the heterologous antigen in vitro to minimize the selective pressure for plasmid loss, while maintaining high levels of expression in vivo to induce an immune response. The three V. cholerae ivi promoters were previously shown to have low levels of expression in vitro on rich medium, but expression was detected from a single chromosomal copy, using a very different indirect reporter assay (Osorio et al., 2005; Lombardo et al., 2007). Thus, luciferase expression from the plasmid constructs was first assessed in vitro. Expression from all bioluminescent reporter plasmids was initially monitored using a defined minimal medium (MOPS) vs. a rich medium (EZ-Rich). Light output was measured and normalized to OD in a 16-h kinetic study. As expected, the promoterless five- and 60-copy plasmids (Pneg5 and Pneg60) showed low background luminescence in both media (Fig. 1a and b). Strains expressing luciferase under the control of PompC displayed constitutively high expression compared with the promoterless negative control in all media tested (Fig. 1a and b). All strains showed a dramatic decrease in expression upon entering the stationary phase. This phenomenon, observed by several groups, is termed abrupt decline of luciferase activity (ADLA) (Koga et al., 2005; Galluzzi & Karp, 2007). The cause of ADLA seems to be a decrease in the availability of reduced flavin mononucleotide (FMNH2), which is the direct electron donor for the bacterial luciferase reaction.

The arginine biosynthesis pathway is well characterized in E. coli (reviewed in Cunin et al., 1986; Charlier & Glansdorff, 2004). The argCBH operon, encoding proteins involved in arginine biosynthesis, is negatively regulated by the ArgR repressor, which binds to ARG boxes within the promoter sequence. Additionally, like many other amino acid biosynthesis genes, the arginine regulon is positively regulated by the stringent response, a multifaceted physiological response in the face of starvation (reviewed in Cunin et al., 1986; Cashel et al., 2009). Thus the full gamut of arginine regulation in V. cholerae can be predicted as follows: full expression in the absence of arginine and under starvation conditions, intermediate expression in the presence of arginine and under starvation conditions, and complete repression in the presence of arginine in rich medium. Consistent with this model, in our kinetics experiments, we saw the highest luciferase activity in both PargC-lux5 and PargC-lux60 in minimal medium (Fig. 1a). In additional experiments, cultures were initially grown in rich medium, pelleted and resuspended in various media. For the five-copy plasmid, repression is very quickly alleviated by growth in minimal medium, but is maintained with as little as 0.5 mM arginine in minimal media (Fig. 2a), whereas for the 60-copy plasmid, addition of the same amount of arginine only partially repressed bioluminescence (Fig. 2b). Increasing the amount of arginine to 5 mM did not increase the repression (data not shown). However, full repression of the 60-copy plasmid was achieved in rich medium (Fig. 2b). These data confirm previous findings that V. cholerae PargC exhibits low expression in vitro.

The fhuCDB operon is involved in ferrichrome iron utilization and is part of the Fur regulon (Rogers et al., 2000). At high concentrations in the bacterium, iron binds to Fur, which then binds to Fur boxes in the promoter regions of iron-regulated genes and prevents transcription. The consensus sequence of the Fur box in V. cholerae is similar to that of E. coli (Mey et al., 2005). In our experiments, the PfhuC promoter was only examined using the five-copy replicons, and showed relatively constant expression in vitro in minimal and rich media that was higher than the promoterless negative control strain (Fig. 1a and b). Kinetic experiments in other media showed high expression in minimal medium containing 50 µM 2,2′-dipyridyl, which limits iron availability, and low expression in minimal medium containing 50 µM FeSO4 (data not shown). Additionally, in induction experiments, despite high basal level expression in rich LB medium, PfhuC was repressed by the addition of 50 µM FeSO4 to minimal medium (Fig. 2c). The design of the IVET study allowed for the inclusion of promoters with low activity in vitro, with higher expression in vivo (Lombardo et al., 2007). Because we are studying the promoter in a multicopy plasmid instead of the single chromosomal copy identified in IVET, we are more likely to see increased activity in vitro.

VCA1008 is a putative outer membrane protein that is necessary for infection of suckling mice (Osorio et al., 2004). The precise regulation of VCA1008 is unknown. Previous studies showed that vca1008 is not induced in vitro in either minimal M9 or rich LB medium, but was induced in vivo in suckling mice using IVET techniques (Osorio et al., 2004). Our initial kinetic studies confirm these findings. Regardless of plasmid copy number or growth conditions, in vitro expression of bioluminescence from Pvca1008 was not detectable above promoterless control levels (Fig. 1a and b). Expression could not be demonstrated in any other media tested, including M9 minimal medium, LB, DMEM tissue culture medium (low and high glucose), or under ToxR-permissive conditions (LB, pH 6.5, at 30 °C) (data not shown). Additionally, expression could not be seen after infection of T84 intestinal cells (data not shown).

In vitro plasmid stability

Expression of heterologous proteins at high levels can decrease plasmid stability (Londono et al., 1996). Luciferase-expressing plasmids have had limited use in vivo until recently because of issues with stability. The pGEN-luxCDABE plasmid was shown by Lane (2007) to be relatively stable in UPEC in vitro for up to 3 days with c. 13% plasmid loss. We examined the in vitro stability of our plasmids in CVD 103-HgR, expressing the lux operon from pGEN-luxCDABE under the transcriptional control of PargC, PfhuC, Pvca1008 and PompC from high- or low-copy plasmids. For live vectors passaged in liquid medium in the absence of antibiotic selection, all the five-copy expression plasmids maintained 100% stability over a 5-day period, whereas the 60-copy plasmids were quickly lost after 2 days (Fig. 3). Plasmid loss did not appear to correlate with high luciferase expression because all three 60-copy plasmids (no promoter, PargC and Pvca1008) express minimal levels of luciferase, while some of the five-copy plasmids exhibit high levels of expression in rich media, including PfhuC and PompC (Fig. 1b). We therefore attribute plasmid loss to high copy number, which has previously been suggested to decrease the fitness of the bacterial vector strain due to the greater amount of energy required to maintain the DNA within the host cell (Glick, 1995).

Figure 3

Plasmid maintenance in the absence of selective pressure in vitro. Strains were serially passaged at 1 : 100 dilutions over 5 days in rich medium without antibiotics. The first time point (0 h) is the first plating after 24 h of growth in liquid medium without selection. The symbols for five-copy plasmids are solid and those for 60-copy plasmids are open.

Evaluating in vivo promoter induction

Suckling mice provide a model for colonization with V. cholerae strains, but their immune systems are too immature to examine the immunogenicity of vaccine strains (reviewed in Klose, 2000). Several groups have utilized various adult mouse models to study the immunogenicity of V. cholerae live vector vaccines with inconsistent results (Butterton et al., 1996; Chen et al., 1998). In this study, we focused only on quantifying in vivo induction of ivi promoters controlling luciferase expression in suckling mice. Because the suckling mice are removed from their mothers, colonization studies are limited to 24 h. We compared expression of the promoters at an early (4 h) and late (24 h) time point.

To study the induction of PargC, PfhuC, Pvca1008 and PompC in vivo, we inoculated 5-day-old CD-1 mice with c. 1 × 107 CFU of each strain and imaged the mice at 0, 4 and 24 h postinoculation using an IVIS 200 imaging system. Mice were anesthetized and imaged individually for the best results. Representative images from each time point of a single mouse in each group inoculated with five-copy plasmids are shown in Fig. 4. Because the amount of bacteria per mouse was different at each time point, the total flux (photons s−1) was normalized to CFU mL−1 of either the initial inoculum (0 h) or intestinal homogenates (4 and 24 h). All strains tested colonized mice at similar levels (Fig. 5).

Figure 4

Representative images demonstrating in vivo promoter induction from five-copy plasmids. Five-day-old CD-1 mice were inoculated with CVD 103-HgR carrying luciferase-expressing plasmids (dose c. 107). One-minute images were acquired immediately after inoculation (0 h) and at 4 and 24 h postinfection. Note that the images represent raw luciferase activity that is not normalized to CFU mL−1 as in Fig. 6.

Figure 5

Colonization levels in suckling mice. Five-day-old CD-1 mice were inoculated with CVD 103-HgR carrying luciferase-expressing plasmids (dose c. 107). Groups were split into two for enumerating bacteria in the intestine at 4 or 24 h. Each symbol represents a single mouse and the bar represents the mean. The 4-h time point is represented by squares and 24 h by triangles. Closed symbols are used for five-copy replicons and open symbols for 60-copy replicons.

Because the argCBGH operon was identified as being in vivo-induced by IVET studies in both suckling mice and human studies, displaying a 700-fold induction level in the suckling mouse intestine (Osorio et al., 2005), it appeared to be an excellent candidate for in vivo bioluminescence studies. In the suckling mouse intestine, PargC in the five-copy plasmid showed very low expression at the 0-h time point and was induced 36-fold above that of the inocula at 4 h and a further 544-fold above the inocula at 24-h (Fig. 6). The 60-copy plasmid had a higher background bioluminescence at the 0-h time point, but achieved one of the highest total expression levels of all constructs, reaching 42.3 vs. 5.1 photons s−1 CFU−1 mL−1 for the five-copy plasmid at 24 h (Fig. 6). Despite high levels of expression in the mouse intestine, the lack of stability in vitro in the absence of selective pressure (Fig. 3) may ultimately detract from using PargC with 60-copy plasmids for antigen expression in V. cholerae live vectors. The five-copy plasmid, however, shows great promise for use in V. cholerae live vector strains because of its tight regulation and stability in vitro and high induction in the mouse intestine.

Figure 6

Quantified in vivo bioluminescence. Regions of interest were calculated for each mouse using living image 3.0 software, giving a unit of photons s−1. The relative bioluminescence per bacteria was calculated for each promoter construct by normalizing the total flux to CFU mL−1 from either the calculated inocula (0 h) or intestinal homogenates (4 and 24 h). N=7–10 mice. Significant differences in luminescence between the 0 h and 4 or 24 h were determined by two-tailed Student's t-test. *P<0.05. **P<0.005.

The fhuACDB operon was also identified as being in vivo-induced by IVET studies in both suckling mice and human studies, displaying induction late in the infection of suckling mice (Schild et al., 2007). By measuring bioluminescence, we demonstrated that there was substantial induction of PfhuC in the suckling mouse intestine, with a 24-fold increase in expression above the inocula, but the induction was statistically significant only at the 24-h time point (Fig. 6), confirming the previous findings of Schild and colleagues. Although expression of the fhuC promoter can be repressed in vitro by adding FeSO4 to the medium, the basal level of expression is still quite high compared with PargC and Pvca1008, and therefore may not be useful for the expression of possibly toxic antigens in V. cholerae live vectors.

As with PargC and PfhuC, the vca1008 promoter was shown by IVET to be inducible in vivo in both suckling mice and human studies, exhibiting the highest induction of all promoters studied in the suckling mouse model (Osorio et al., 2005). Although we were unable to detect expression of luciferase from the vca1008 promoter in vitro, consistent with previous findings (Osorio et al., 2004), we observed excellent induction in vivo. This promoter is strongly induced in suckling mice using both the five- and 60-copy replicons. Activation of the vca1008 promoter occurred by 4-h postinfection and remained elevated at 24 h (Figs 4 and 6). In additional experiments, expression from Pvca1008 was shown to occur as early as 1 h postinoculation, the earliest time point tested (data not shown). Interestingly, the maximum amount of bioluminescence was equivalent for both the five- and 60-copy plasmids. We conclude that regulation of this promoter is exquisitely sensitive to the in vivo environment, as it exhibits virtually no leakiness of expression in vitro even when present in high copy number, but shows excellent induction in vivo. Pvca1008 represents an intriguing candidate for achieving tightly regulated expression of heterologous antigens in V. cholerae live vector vaccines. In the case of expression of antigens that are toxic to the host strain, maximal repression could be attained in vitro during preparation of the strain for immunizations, possibly ensuring better immunogenicity due to improved fitness of the live vector.

As discussed above, because the E. coli OmpC protein does not have an orthologue in V. cholerae, we were uncertain as to whether PompC would be induced in vivo in V. cholerae. Surprisingly, we observed a 100-fold induction of PompC in the sucking mouse intestine. Indeed, PompC showed the highest total expression of bioluminescence for the five-copy replicons, with an average of 44.8 photons s−1 CFU−1 mL−1, a level equivalent to expression levels from PargC carried by 60-copy plasmids (Fig. 6). As with PargC, use of PompC with five-copy replicons shows great promise for use in V. cholerae live vector strains because of its in vitro stability and high induction in the mouse intestine.

Concluding remarks

In summary, this study has identified three excellent candidate promoters, PargC, Pvca1008, and PompC, for use in regulated expression of heterologous antigens in attenuated V. cholerae live vector vaccine strains. Because the promoters were originally identified in an El Tor strain and then studied here using bioluminescence in a classical strain it is likely that they will function well in both biotypes. The rational identification and direct demonstration of in vivo expression of these promoters has identified useful tools for advancing the development V. cholerae multivalent live vector vaccines.

Our series of luciferase-expressing stabilized plasmids could also be useful for the study of other enteric pathogens. For bacteria with more robust mouse colonization models, strains expressing luxCDABE from the ompC promoter could be very useful for studying colonization dynamics, without having to create chromosomal lux fusions. Our plasmids can be transformed into other bacterial strains, including S. Typhi. After transforming five-copy replicons expressing PompC-controlled luciferase into S. Typhi and inoculating mice intranasally, we were able to track both colonization and tissue distribution over a 5-day period (data not shown). These plasmids could also be used to examine in vivo induction of other predicted in vivo-induced promoters in V. cholerae or other bacteria. More detailed studies of induction kinetics or in situ induction studies could be performed with enteric bacteria because the gastrointestinal tract can be removed from colonized mice and quantitatively imaged. Additionally, bacteria expressing luciferase have been used to target tumor cells and noninvasively follow tumor growth and metastasis over time without the need for exogenous luciferin injections, as is necessary when imaging tumor cells that have been engineered to express the firefly luc gene. Min (2008a, b) have recently used E. coli expressing luciferase from a plasmid stabilized using the asd balanced-lethal host–vector system to image a variety of tumors in living mice. Our system could easily be adapted for such use.


We wish to acknowledge the advice and gift of plasmid constructs from J. Galen. Thanks are due to L. Plemons for excellent technical assistance. A.-M. Hansen, N. Morin and M.-J. Lombardo are also thanked for advice and editorial assistance. We are grateful to M. Lane and H. Mobley for the gift of pGENluxCDABE. This work was supported by R0I AI19716 (J.B.K.). C.E.M. was supported by a predoctoral fellowship on T32 DK067872, ‘Research Training in Gastroenterology’ (J.P. Raufman, PI).


  • Editor: Patrik Bavoil


View Abstract