The 5′ untranslated region-mediated enhancement of intracellular listeriolysin O production is required for Listeria monocytogenes pathogenicity

Authors


Summary

Listeriolysin O (LLO) and ActA are essential virulence determinants for Listeria monocytogenes pathogenesis. Transcription of actA and hly, encoding LLO, is regulated by PrfA and increases dramatically during intracellular infection. The 5′ untranslated regions (5′ UTRs) of actA and prfA have been shown to upregulate expression of their respective gene products. Here, we demonstrate that the hly 5′ UTR plays a critical role in regulating expression of LLO during intracellular infection. Deletion of the hly 5′ UTR, while retaining the hly ribosome binding site, had a moderate effect on LLO production during growth in broth culture, yet resulted in a marked decrease in LLO levels during intracellular infection. The diminished level of LLO resulted in a significant defect in bacterial cell-to-cell spread during intracellular infection and a 10-fold reduction in virulence during in vivo infection of mice. Insertion of the hly 5′ UTR sequence between a heterologous promoter and reporter gene sequences indicated that the hly 5′ UTR functions independent of PrfA-mediated transcription and can enhance expression of cis-associated genes through a mechanism that appears to act at both a post-transcriptional and translational level. The ability of the hly 5′ UTR to increase gene expression can be exploited to achieve PrfA-independent complementation of virulence genes and high-level expression of single copy heterologous genes in L. monocytogenes.

Introduction

Listeria monocytogenes is a facultative intracellular bacterial pathogen of humans and a wide variety of animal species (Vazquez-Boland et al., 2001). Upon entering host cells, expression of many virulence determinants is cell compartment-specific (Bubert et al., 1999). Expression of listeriolysin O (LLO) and PI-PLC, which mediate bacterial escape from phagocytic vacuoles, is activated at the transcriptional level within the vacuole, while expression of ActA and PC-PLC, which mediate actin-based motility and cell-to-cell spread, and escape from secondary spreading vacuoles, respectively, is activated once bacteria access the host cell cytosol (Bubert et al., 1999; Freitag and Jacobs, 1999; Shetron-Rama et al., 2002). Transcription of the genes encoding these principal virulence determinants is controlled by PrfA, a transcriptional activator that binds to 14 bp palindromic sequences present upstream of virulence gene promoters (Leimeister-Wachter et al., 1990; Chakraborty et al., 1992; Sheehan et al., 1995; Milohanic et al., 2003). During intracellular infection, transcription of hly, encoding LLO, is upregulated ∼20-fold (Moors et al., 1999b). This induction of LLO expression is likely required for efficient secondary vacuole escape, as bacterial strains that produce decreased amounts of LLO during intracellular infection display defects in cell-to-cell spread (Dancz et al., 2002).

ActA is one of the most abundant L. monocytogenes proteins produced during intracellular infection (Brundage et al., 1993), with ActA protein levels being regulated by transcriptional and post-translational mechanisms (Brundage et al., 1993; Moors et al., 1999a,b). Transcription of actA is induced ∼150- to 200-fold during infection of host cells (Moors et al., 1999b; Shetron-Rama et al., 2002), and small decreases in ActA protein levels dramatically alter the efficiency of cell-to-cell spread by decreasing the frequency of actin-tail formation (Smith et al., 1996; Moors et al., 1999a; Wong et al., 2004). However, overexpression of ActA also appears to impair cell-to-cell spread (Lauer et al., 2002). Similarly, increased levels of cytosolic LLO are toxic to host cells (Higgins et al., 1999; Decatur and Portnoy, 2000). Thus, stringent regulation of virulence gene expression is critical for productive intracellular infection by L. monocytogenes.

Most studies of virulence gene regulation in L. monocytogenes have focused on the contribution of PrfA in activating transcription. However, several studies have demonstrated that additional factors regulate expression of L. monocytogenes virulence determinants during intracellular infection (Moors et al., 1999b; Shetron-Rama et al., 2002; 2003). The 5′ untranslated regions (UTRs) of prfA and actA transcripts have been shown to regulate expression of their cognate gene products. The prfA 5′ UTR functions as a thermosensor to regulate translation of PrfA. At 30°C, the prfA 5′ UTR adopts a secondary structure that prevents initiation of translation on prfA transcripts. At 37°C, this inhibitory structure is no longer energetically favourable, thereby allowing translation of prfA transcripts (Johansson et al., 2002). Similarly, the actA 5′ UTR functions to maximize expression of ActA (Wong et al., 2004). Deletion of 128 nucleotides of the actA 5′ UTR, while retaining the actA ribosome binding site (RBS), results in a fourfold reduction in intracellular ActA levels and an inability to polymerize host actin (Wong et al., 2004). Although the mechanism underlying this function remains to be determined, the secondary structure of the actA 5′ UTR is believed to play a role in regulating ActA expression.

In this study, we examined the role of the hly 5′ UTR in regulating expression of LLO during extracellular growth and intracellular infection. Deletion of the hly 5′ UTR, while retaining the hly RBS, yielded a moderate change in LLO levels during growth in broth culture, yet resulted in a marked decrease in LLO production during intracellular infection. The diminished levels of LLO produced by the hly 5′ UTR mutant during cytosolic growth correlated with a significant defect in bacterial cell-to-cell spread in tissue culture cells and a 10-fold decrease in virulence during in  vivo infection of mice. Fusion of the hly 5′ UTR to a heterologous promoter and reporter gene sequences resulted in enhanced gene expression, indicating that the hly 5′ UTR can function in cis independent of PrfA-mediated activation. Lastly, we demonstrate that the ability of the hly 5′ UTR to increase gene expression can be exploited to achieve PrfA-independent complementation of L. monocytogenes virulence genes and yield high-level expression of heterologous genes in single copy.

Results

The hly 5′ UTR is required for maximal LLO production during intracellular infection

Prior studies have shown that the actA 5′ UTR functions in regulating ActA expression and is required for full virulence of L. monocytogenes (Wong et al., 2004). Given the similarities in induction of intracellular expression and the importance of ActA and LLO for L. monocytogenes pathogenesis (Moors et al., 1999b; Dancz et al., 2002; Wong et al., 2004), we sought to determine whether the hly 5′ UTR regulates expression of LLO. To determine the importance of the hly 5′ UTR, deletions within the hly 5′ UTR sequence were introduced into the chromosome of wild-type L. monocytogenes strain 10403S. hly is predominantly transcribed from a PrfA-dependent promoter (Phly) (Domann et al., 1993). The PrfA-dependent promoter initiates transcription at two sites, P2 and P1, which have been mapped by primer extension to produce transcripts with a 5′ UTR of 133 and 122 nucleotides respectively (Mengaud et al., 1989) (Fig. 1A). A total of 101 bp of the hly 5′ UTR sequence were deleted from the hly locus to yield strain DH-L1231, which retains the PrfA-dependent Phly promoter, the P2 and P1 transcription initiation sites, and the RBS (Fig. 1B).

Figure 1.

Construction of a L. monocytogenes strain containing a deletion of the hly upstream regionA. Detailed depiction of the hly upstream region. The PrfA-binding site (PrfA box) for the hly promoter is boxed, the -10 promoter element is underlined, and the two transcription initiation sites, P2 and P1, are indicated by +1 and +12 respectively. The hly ribosome binding site (RBS) is shaded, while the open block arrow denotes the hly coding region. The regions enclosed by brackets indicate nucleotides deleted from L. monocytogenes strain DH-L1231B. Schematic representation of L. monocytogenes hly 5′ UTR mutant strain, DH-L1231. Dark arrows denote the PrfA-dependent promoters for hly (Phly) and plcA (PplcA); filled rectangles, open block arrows, dashed lines, and open ovals denote the PrfA-binding site, hly coding region, hly 5′ UTR sequence, and hly RBS respectively.

The effect of deleting the hly 5′ UTR on LLO production was assessed during growth in broth culture and during intracellular infection. During growth in brain heart infusion (BHI) broth, no significant difference in bacterial growth rate was observed (data not shown), while a slight decrease (∼31%) in LLO haemolytic activity present in culture supernatants was detected from strain DH-L1231 [92 ± 25 haemolytic units (HU)] relative to wild-type 10403S (134 ± 40 HU) (Fig. 2A). LLO protein levels in bacterial cell pellets and supernatant fractions of DH-L1231 cultures were also mildly diminished when compared with 10403S (Fig. 2B, BHI). Primer extension analysis indicated that hly transcript levels were 3.5-fold (71%) lower in DH-L1231 compared with wild-type 10403S and confirmed that transcription of hly initiated at the predicted P2 transcription initiation site of strain DH-L1231 (Fig. 2C) However, no hly transcripts were detected that initiated at the P1 transcription initiation site. This may indicate that the hly P1 transcription initiation site in 10403S is actually an mRNA processing site within hly transcripts that initiate at P2 or that DNA sequences encompassing the hly 5′ UTR are required for PrfA-dependent transcription initiation at the P1 site. Northern blot analysis of hly transcripts yielded a similar decrease in hly transcript levels in the hly 5′ UTR mutant relative to wild-type 10403S (data not shown). Thus, deletion of the hly 5′ UTR resulted in a moderate decrease in LLO production and a more severe decrease in hly transcript levels during growth in broth culture.

Figure 2.

Expression of LLO during growth in BHI brothA. Haemolytic activity of L. monocytogenes strains. Fourteen to 16 h cultures of wild type (10403S), LLO-negative (DP-L2161) and the hly 5′ UTR mutant (DH-L1231) were diluted 1:10 in BHI or BHIC broth and grown for 5 h at 37°C. Haemolytic activity (HU) present in culture supernatants was determined as described in Experimental procedures. The means and standard deviations of three independent experiments are shown. Haemolytic activity for DP-L2161 culture supernatants was not detectedB. Western blot analysis of LLO expression. Aliquots of cultures used in A were centrifuged. Cell pellets (CP) were resuspended and digested with lysozyme, and supernatants (SN) were precipitated for 1 h on ice in the presence of 10% trichloroacetic acid (v/v). Samples were centrifuged and resuspended in protein sample buffer. Protein samples from a culture volume equivalent to 100 µl (CP) or 50 µl (SN) of an OD600 = 1.5 were separated by SDS-PAGE and immunoblotted using a rabbit polyclonal anti-LLO antibody. Numerical values indicate the percent band intensity relative to 10403S as determined by densitometry analysis and are representative of at least three independent experimentsC. Primer extension analysis of hly transcripts. Total RNA was isolated from bacterial strains used in A that had been grown in BHI broth. 7.5 µg of total RNA was hybridized to a 32P-end-labelled primer specific for hly. Primer extension was performed as described in Experimental procedures. A radiolabelled primer specific to the constitutively transcribed iap gene (Kohler et al., 1990) was simultaneously hybridized to total RNA as an internal load control. Identities of extended products are indicated. The P1 and P2 products correspond to hly transcripts resulting from PrfA activation of the Phly promoter in 10403S. Lane M is a radiolabelled size ladder. Band intensities of probe-associated transcripts were quantified by phosphorimager analysis. Numerical values indicate the percent hly/iap transcript relative to 10403S and represent the mean and standard deviation of three independent experiments.

Despite a modest effect of the hly 5′ UTR on LLO levels during growth in BHI broth, we considered the possibility that deletion of the hly 5′ UTR might result in a greater defect in LLO production when hly expression is upregulated by PrfA. Treatment of BHI broth with activated charcoal (BHIC) results in activation of PrfA by chelating a compound that inhibits the activity of PrfA (Ermolaeva et al., 2004). Thus, during growth in BHIC, a 10- to 20-fold increase in LLO levels is observed relative to the levels produced in BHI broth (Geoffroy et al., 1987; Ripio et al., 1996). Although levels of LLO production were increased compared with growth in BHI, growth of the hly 5′ UTR mutant in BHIC resulted in a similar decrease (∼37%) in haemolytic activity of DH-L1231 (982 ± 36 HU) relative to 10403S (1554 ± 192 HU) (Fig. 2A), without affecting bacterial growth rate (data not shown). Diminished LLO levels relative to 10403S were also observed by Western blot analysis of bacterial cell pellets and supernatant fractions of DH-L1231 (Fig. 2B, BHIC). Thus, deletion of the hly 5′ UTR resulted in a similar defect in LLO production during growth in PrfA-activating conditions.

As PrfA-mediated activation of virulence genes is maximal during intracellular infection, we determined the effect of deleting the hly 5′ UTR on LLO production during growth inside J774 host cells. Approximately 55% less secreted LLO relative to wild-type 10403S was immunoprecipitated from metabolically labelled J774 cells during infection with the hly 5′ UTR mutant (Fig. 3A). A greater reduction in bacterial-associated LLO levels (78%) relative to 10403S was also observed during infection of J774 cells with DH-L1231 (Fig. 3B). The decrease in LLO production by the hly 5′ UTR mutant relative to wild type could not be attributed to differences in intracellular growth, as DH-L1231 bacteria were recovered in similar numbers as 10403S bacteria during infection of J774 cells (data not shown).

Figure 3.

The hly 5′ UTR is required for maximal LLO production during intracellular infectionA. Immunoprecipitation of LLO from J774 cells infected with L. monocytogenes. 10403S and hly 5′ UTR mutant, DH-L1231, were grown in BHI broth and used to infect J774 cells as described in Experimental procedures. At 6 h post infection, bacterial proteins were metabolically labelled for 1 h with [35S]-methionine, and LLO was immunoprecipitated from host cell lysates using monoclonal anti-LLO antibody (B3-19). Immunoprecipitated proteins (IP) were resuspended in 2× FSB and resolved by SDS-PAGE. The resulting gel was exposed to a phosphorimaging screen for 7 days and subjected to phosphorimager analyses. Numerical values indicate the percent band intensity of LLO relative to 10403S and represent the mean and standard deviation of three independent experiments. As previously shown (Moors et al., 1999b), two distinct species of LLO protein were detected in immunoprecipitated fractionsB. Western blot analysis of bacterial-associated LLO during infection of J774 cells. L. monocytogenes strains described in A were grown in BHI broth and used to infect monolayers of J774 cells as described in Experimental procedures. At 7 h post infection, bacteria were pelleted from host cell lysates. Bacterial cell pellets (CP) were resuspended, digested with lysosome, and 2× FSB was added. Protein samples equivalent to 2 × 108 cfu were separated by SDS-PAGE and immunoblotted using monoclonal anti-LLO antibody (B3-19). Numerical values indicate the percent band intensity relative to 10403S as determined by densitometry analysis and represent the mean and standard deviation of three independent experimentsC. Primer extension analysis of hly transcripts. A parallel infection to that described in B was used to harvest RNA from bacteria isolated from J774 cell lysates. RNA was also isolated from 10403S and DH-L1231 bacteria grown in BHI broth culture as described in Fig. 2A. Three micrograms of total RNA isolated from intracellular bacteria, and 15 µg of total RNA isolated from bacteria grown in BHI broth were hybridized to a 32P-end-labelled primer specific for hly. Primer extension was performed as described in Experimental procedures. A radiolabelled primer specific to the plcA gene, which is divergently transcribed from hly, was simultaneously hybridized to total RNA as an internal load control. Identities of extended products are indicated. The P1 and P2 products correspond to hly transcripts that result from PrfA activation of the Phly promoter in 10403S. Band intensities of probe-associated transcripts were quantified by phosphorimager analysis. Numerical values indicate the percent of hly/plcA transcript relative to 10403S and represent the means and standard deviations of three independent experiments.

To determine whether the decrease in LLO levels detected from the hly 5′ UTR mutant relative to wild type during intracellular infection correlated with a decrease in hly transcript levels, RNA was harvested from 10403S and DH-L1231 bacteria grown inside J774 cells, and hly transcripts were detected directly by primer extension. During intracellular infection, DH-L1231 produced approximately fourfold less hly transcripts than 10403S; a similar 3.5-fold decrease in hly transcript levels in the hly 5′ UTR mutant relative to wild type was observed during growth in BHI broth (Figs 2C and 3C). Furthermore, upregulation of Phly during intracellular infection did not seem to be impaired in DH-L1231, as an ∼30-fold increase in hly transcription was observed in both 10403S and DH-L1231 upon growth in the cytosol relative to growth in BHI broth. It should be noted that fivefold more RNA was used in the extension reactions for samples isolated from bacteria grown in BHI broth than in reactions used with samples harvested from intracellular bacteria (Fig. 3C). Importantly, upregulation of the plcA promoter was unaffected by deletion of the hly 5′ UTR, despite sharing the same PrfA-box as Phly (Fig. 1B). The observed ∼30-fold induction of Phly during intracellular infection corresponds well with the ∼20-fold increase in hly promoter activation measured under similar conditions using transcriptional fusions (Moors et al., 1999b). Taken together, these results suggest that the hly 5′ UTR sequence is dispensable for PrfA-mediated induction of Phly during intracellular infection, but is required for maximal LLO production during growth inside host cells.

The hly 5′ UTR is required for efficient cell-to-cell spread and virulence of L. monocytogenes

We next examined whether the decrease in intracellular LLO levels observed upon deletion of the hly 5′ UTR translated into a defect in cell-to-cell spread and virulence of L. monocytogenes. Deletion of the hly 5′ UTR had little effect on intracellular growth of L. monocytogenes, because DH-L1231 grew similarly to wild-type 10403S during infection of murine bone marrow-derived macrophages (BMM) (Fig. 4A) and in J774 cells (data not shown). Furthermore, microscopic analysis of infected BMM did not indicate any significant differences in intracellular infection over the 12-h period examined in Fig. 4A. In contrast, plaquing analysis over a 72 h infection period in murine L2 fibroblasts indicated a deficiency in plaque formation for the hly 5′ UTR deletion mutant, as DH-L1231 yielded an average plaque size 68 ± 8% of that formed by 10403S (Fig. 4B). Prior studies from our group have shown that decreases in intracellular LLO production result in concomitant decreases in plaque size, which most likely results from a deficiency in secondary vacuole escape following cell-to-cell spread (Dancz et al., 2002). Thus, the hly 5′ UTR likely plays a role in regulating intracellular LLO production to obtain protein levels necessary for efficient cell-to-cell spread.

Figure 4.

The hly 5′ UTR is required for efficient cell-to-cell spread during intracellular infectionA. Intracellular growth of L. monocytogenes in murine bone marrow-derived macrophages (BMM). Monolayers of BMM seeded onto glass coverslips were infected with L. monocytogenes strains at an moi of 1:3 as described in Experimental procedures. At the indicated times post infection, coverslips were removed and the number of intracellular bacteria determined. Data represent one of three independent experiments performed in triplicate with similar resultsB. Plaque formation in L2 fibroblasts. Wild type (10403S), LLO-negative (DP-L2161), or the hly 5′ UTR mutant (DH-L1231) were grown in BHI broth and added to monolayers of mouse L2 fibroblasts for 1 h. The infected monolayers were washed with PBS, and a medium-agarose overlay containing gentamicin was added to kill extracellular bacteria. Intracellular growth and cell-to-cell spread of bacteria were visualized after 72 h by the formation of clearing zones (plaques) within the L2 monolayers. The diameters of 15 plaques/sample were measured. Values are expressed as the percent diameter of plaques relative to 10403S and represent the means and standard deviations of three independent experiments. n.d., no plaques detected.

To assess whether the defects in cell-to-cell spread associated with deletions in the hly 5′ UTR would correlate with a virulence defect in vivo, we determined the LD50 value for DH-L1231 during infection of BALB/c mice. The LD50 for wild-type 10403S was 1–3 × 104, whereas the LD50 for DH-L1231 was 2–3 × 105, indicating that deletions within the hly 5′ UTR result in a 10-fold decrease in virulence. Taken together, these results suggest that the hly 5′ UTR is required for maximal expression of LLO during intracellular infection, which in turn is necessary for full virulence of L. monocytogenes.

The hly 5′ UTR regulates LLO production independent of PrfA-mediated transcription of hly

While deletions within the hly 5′ UTR had a detectable effect on LLO expression in BHI broth (Fig. 2A and 2B), data in Fig. 3 indicated that the hly 5′ UTR is required for maximal LLO expression under intracellular growth conditions where PrfA activation of hly expression is greatest. Although our results indicated that deletion of the hly 5′ UTR did not affect upregulation of Phly during intracellular infection, hly transcripts were reduced in the hly 5′ UTR mutant relative to wild type regardless of the growth condition (Fig. 3C). This raised the possibility that PrfA-mediated transcription initiation from Phly is dependent on DNA sequences within the hly 5′ UTR, especially given the absence of detectable transcript initiating from the P1 promoter in DH-L1231 (Figs 2C and 3C). To address this possibility, we uncoupled hly transcription from PrfA-dependent promoter activation and examined the effect of deleting the hly 5′ UTR on expression of LLO. Transcription of hly was placed under the control of the constitutive HyperSPO1 promoter (Quisel et al., 2001) (Fig. 5A), and hly constructs harbouring varying lengths of the hly 5′ UTR were expressed in DP-L2161, a 10403S-derived strain containing a deletion of the hly gene (Jones and Portnoy, 1994). Strain DH-L911 harboured the entire hly 5′ UTR (Fig. 5A) and was used as the reference strain by which to compare the effect of deletions of the hly 5′ UTR on LLO production. Progressive deletions of 34, 68 and 113 bp from the 5′ end of the hly upstream region were used to generate strains DH-L934, DH-L935 and DH-L936 respectively (Fig. 5A). The hly expression constructs were integrated in single copy within the ectopic tRNAArg locus (Lauer et al., 2002) of DP-L2161.

Figure 5.

The hly 5′ UTR can enhance expression of LLO independent of PrfA-mediated transcription of hlyA. Schematic representation of HyperSPO1 promoter-controlled hly strains harbouring various lengths of the hly 5′ UTR sequence. The dark arrow, dashed line, oval and open block arrow denote the HyperSPO1 promoter, hly 5′ UTR sequence, hly RBS and hly coding region respectively. The regions deleted within the hly 5′ UTR are given relative to the P2 transcription initiation site of hly (Fig. 1A)B. Haemolytic activity of L. monocytogenes strains. Fourteen to 16 h cultures of wild type (10403S), LLO-negative (DP-L2161) and the HyperSPO1 promoter-controlled hly strains were diluted 1:10 in BHI broth and grown for 5 h at 37°C. Haemolytic activity present in culture supernatants was determined as described in Experimental procedures. The means and standard deviations of three independent experiments are shownC. Quantification of LLO expression by Western blot. Aliquots of cultures used in B were centrifuged, and the cell pellets (CP) were resuspended and digested with lysozyme. Culture supernatants (SN) were precipitated for 1 h on ice in the presence of 10% trichloroacetic acid (v/v). Protein samples from a culture volume equivalent to 100 µl (CP) or 50 µl (SN) of an OD600 = 1.5 were separated by SDS-PAGE and immunoblotted with a rabbit polyclonal anti-LLO antibody. Protein band intensities were quantified by densitometry. Data shown are representative of at least three independent experimentsD. Primer extension analysis of hly transcript levels. 7.5 µg of total RNA from the HyperSPO1 promoter-controlled hly strains was hybridized to a 32P-end-labelled primer specific for hly. Primer extension was performed as described in Experimental procedures. A radiolabelled primer specific to the iap gene was simultaneously hybridized to total RNA as an internal load control. Identities of extended products are indicated. Band intensities of probe-associated transcripts were quantified by phosphorimager analysis. Numerical values indicate the percent hly/iap transcript relative to DH-L911 and represent the means and standard deviations of three independent experiments.

To determine the effect of progressive deletion of the hly 5′ UTR on LLO production, we measured the haemolytic activity present in culture supernatants of each strain. Decreasing the length of the hly 5′ UTR from the 5′ end resulted in concomitant decreases in haemolytic activity. While the haemolytic activity for all of the HyperSPO1 promoter-controlled hly strains was greater than that observed for wild-type 10403S, the haemolytic activity of strains DH-L934, DH-L935 and DH-L936 was 1.2-, 1.6- and 2.2-fold less, respectively, than the haemolytic activity of DH-L911 (Fig. 5B). Western blot analysis of LLO protein present in bacterial cell pellets and supernatant fractions correlated with the observed haemolytic activities for each of the strains examined (Fig. 5C). Given that each progressive truncation of the hly 5′ UTR decreased LLO production, these results suggested that sequences from +1 through +113 of the hly 5′ UTR play a role in enhancing expression of LLO and can function independent of PrfA-mediated transcription initiation.

We next examined whether the reduced LLO levels observed upon deletion of the hly 5′ UTR correlated with a decrease in hly transcript levels. Primer extension analysis indicated that a general decrease in hly transcripts was observed upon truncation of the hly 5′ UTR (Fig. 5D), and confirmed that transcription of hly initiated at the predicted sites within the HyperSPO1 promoter-controlled hly strains (Fig. 5A). DH-L934 and DH-L935 generated, respectively, 1.4-fold and 2.3-fold less hly transcripts than DH-L911. Similar results were obtained by Northern blot analysis (data not shown). However, a strict correlation between hly transcript and LLO protein levels was not observed. Although DH-L936 produced less LLO protein and haemolytic activity than DH-L935 (Fig. 5B and C), DH-L936 produced slightly more hly transcripts than DH-L935 (Fig. 5D). Thus, the data in Fig. 5 indicated that the hly 5′ UTR can mediate enhanced production of LLO independent of PrfA-regulated transcription of hly.

The hly 5′ UTR can enhance expression of a heterologous gene product in L. monocytogenes

Although the hly 5′ UTR could enhance expression of LLO independent of PrfA-mediated transcription of hly (Fig. 5), it is possible that sequences within the hly coding region may be required for hly 5′ UTR-mediated enhancement of LLO expression. Alternatively, the hly 5′ UTR may be sufficient to indiscriminately enhance expression of cis-associated gene products. To distinguish between these possibilities, we generated a L. monocytogenes strain in which the gfp gene encoding green fluorescent protein (GFP) from Aequorea victoria was fused directly downstream of the hly 5′ UTR and transcribed from the HyperSPO1 promoter. The full-length hly 5′ UTR-gfp fusion was integrated in single copy within the tRNAArg locus of 10403S to yield strain DH-L1039. Additional constructs harbouring deletions within the hly 5′ UTR were constructed and placed under transcriptional control of the HyperSPO1 promoter to yield progressive deletions of the hly 5′ UTR sequence fused to gfp, generating strains DH-L1118, DH-L1119 and DH-L1041 respectively (Fig. 6A). The effect of progressive deletions of the 5′ end of the hly 5′ UTR on GFP expression was determined. Using Western blot analysis, fusion of the complete hly 5′ UTR to gfp (strain DH-L1039) resulted in a significant increase in GFP expression levels compared with strain DH-L1041, which harbours a 113 bp deletion of the hly 5′ UTR sequence (Fig. 6B). The increase in GFP protein correlated with a substantial increase in fluorescence intensity of bacteria as determined by fluorimetric analysis. Cultures of DH-L1039 generated a 43-fold higher fluorescence intensity (89 ± 12 units) compared with DH-L1041 (2 ± 2 units) (Fig. 6C). Progressive truncations from the 5′ end of the hly 5′ UTR resulted in concomitant decreases in GFP protein and associated fluorescence of bacteria. Fluorescence intensity of DH-L1118 (47 ± 9 units) and DH-L1119 (21 ± 4 units) were decreased approximately two- and fourfold, respectively, compared with DH-L1039 (89 ± 12 units) (Fig. 6C).

Figure 6.

hly 5′ UTR-mediated enhancement of GFP expressionA. Schematic representation of HyperSPO1 promoter-controlled gfp strains harbouring various lengths of the hly 5′ UTR sequence. The dark arrow, dashed line, oval and open block arrow denote the HyperSPO1 promoter, hly 5′ UTR sequence, hly RBS and gfpmut2 coding region respectively. The regions deleted within the hly 5′ UTR are indicated relative to the P2 transcription initiation site of hlyB. Western blot analysis of GFP expression. Fourteen to 16 h cultures of wild type (10403S) and the HyperSPO1 promoter-controlled gfp strains were diluted 1:10 in BHI broth and grown for 3 h at 37°C. Aliquots of cultures were centrifuged and cell pellets were resuspended and digested with lysozyme. Protein samples from a culture volume equivalent to 100 µl of an OD600 = 1.0 were separated by SDS-PAGE and analysed by Western blot using a rabbit polyclonal anti-GFP antibodyC. Fluorimetric analysis of GFP expression. Aliquots of cultures used in B were pelleted and resuspended in PBS. Relative fluorescence units of cultures was measured as described in Experimental procedures and normalized against OD600 (RFU/OD600)D. Primer extension analysis of gfp transcript levels. Total RNA was isolated from aliquots of cultures used in B. 7.5 µg of RNA was hybridized to a 32P-end-labelled primer specific for gfp. Primer extension was performed as described in Experimental procedures. A radiolabelled primer specific to iap was simultaneously hybridized to total RNA as an internal load control. Identities of extended products are indicated. Lane M is a radiolabelled size ladder. Band intensities of extended transcripts were quantified by phosphorimager analysis. Numerical values indicate the percent gfp/iap transcript relative to DH-L1039 and represent the means and standard deviations of three independent experiments.

Primer extension analysis of gfp transcripts indicated that progressive truncation from the 5′ end of the hly 5′ UTR resulted in a general decrease in gfp transcript levels (Fig. 6D), while confirming that transcription of gfp initiated at the predicted sites (Fig. 6A). However, gfp transcript levels were not directly proportional to GFP protein levels. Whereas strain DH-L1119 produced more GFP protein and 10-fold higher fluorescence than DH-L1041 (Fig. 6B and C), gfp transcript levels remained similar within the two strains (Fig. 6D). Furthermore, while fluorescence intensity increased ∼43-fold, respectively, in DH-L1039 compared with DH-L1041 (Fig. 6C), gfp transcript levels were only increased ∼2.5-fold in DH-L1039 compared with DH-L1041 (Fig. 6D). This result suggests that the mechanism responsible for the significant enhancement of GFP protein production mediated by the hly 5′ UTR does not predominantly involve an increase in transcript levels. Fusion of full-length and truncated hly 5′ UTR fragments to the reporter genes prfA and gus, which encodes β-glucuronidase, yielded similar results. Significant decreases in protein levels were observed upon progressive deletion of the hly 5′ UTR from the 5′ end, while decreases in transcript levels were not directly proportional to the decrease in protein levels (data not shown). Collectively, these results indicated that the hly 5′ UTR can function in cis to enhance expression of a heterologous gene product through a mechanism that predominantly functions at a post-transcriptional level.

Discussion

Previous studies have implicated the 5′ UTRs of actA and prfA in regulating expression of their respective gene products during intracellular infection by L. monocytogenes (Johansson et al., 2002; Wong et al., 2004). In addition, a recent report has determined that the 5′ UTR of inlAB functions at a post-transcriptional level to control production of InlA and InlB, which mediate L. monocytogenes invasion of non-professional phagocytic host cells (Stritzker et al., 2005). In our current study, we have demonstrated a role for the hly 5′ UTR in modulating LLO expression levels. Although deletion of the hly 5′ UTR had a moderate effect on LLO production during growth in broth culture, our results indicate that the hly 5′ UTR naturally functions to enhance LLO production during intracellular growth to levels required for full virulence of L. monocytogenes during in vivo infection. During infection of J774 cells, the hly 5′ UTR mutant, DH-L1231, which lacks the majority of the hly 5′ UTR, but retains the hly RBS, produced ∼55% less secreted LLO and ∼78% less bacterial-associated LLO relative to wild-type 10403S (Fig. 3A and B). This decrease in intracellular LLO production correlated with a 32% defect in plaque size formation during infection of L2 fibroblasts (Fig. 4B) and a 10-fold reduction in virulence in a mouse model of infection. Prior studies from our group have demonstrated that decreases in production of LLO during intracellular infection result in concomitant decreases in plaque size during infection of L2 fibroblasts (Dancz et al., 2002). As no defect in intracellular growth rate was observed for strains harbouring deletions of the hly 5′ UTR sequence (Fig. 4A), our data suggest that intracellular LLO levels produced by a strain harbouring a deletion within the hly 5′ UTR are inadequate to allow efficient secondary vacuolar escape subsequent to cell-to-cell spread.

The two- to fivefold reduced levels of LLO observed in the hly 5′ UTR deletion mutant relative to wild type during intracellular growth correlated with an approximately fourfold decrease in hly transcript levels in the DH-L1231 mutant strain compared with wild type (Fig. 3C). During growth in BHI broth, hly transcript levels were also reduced by ∼3.5-fold in DH-L1231 relative to 10403S (Figs 2C and 3C), although LLO levels were only mildly diminished (1.2- to 1.6-fold) in the hly 5′ UTR mutant compared with wild type (Fig. 2B). As hly transcription increases significantly upon intracellular growth (Moors et al., 1999b), these results may indicate that the hly 5′ UTR increases LLO production slightly for each hly transcript produced, perhaps by stabilizing hly mRNA from degradation, and that upon upregulation of hly transcription, the effect of the hly 5′ UTR becomes amplified to result in marked changes in intracellular LLO levels. Alternatively, transcripts containing the full-length hly 5′ UTR may be more efficiently translated during intracellular infection relative to extracellular growth.

Although deletion of the hly 5′ UTR did not appear to affect PrfA-mediated activation of Phly during intracellular infection (Fig. 3C), it remained possible that the DNA sequences constituting the hly 5′ UTR might participate in PrfA-dependent activation of the hly promoter. Therefore, we uncoupled transcription of hly from PrfA-dependence and placed transcription under the control of the constitutive HyperSPO1 promoter. Deletion of 113 bp of the hly upstream region within DH-L936 resulted in a twofold decrease in LLO protein and haemolytic activity produced in broth culture relative to DH-L911, a HyperSPO1 promoter-controlled hly strain containing the entire hly 5′ UTR sequence (Fig. 5). Furthermore, truncations from the 5′ end of the hly 5′ UTR resulted in concomitant decreases in LLO production (Fig. 5), indicating that sequences along the entire hly 5′ UTR are involved in regulating LLO expression when hly is transcribed from the HyperSPO1 promoter. Because initiation of transcription from the HyperSPO1 promoter is presumably unaffected by hly 5′ UTR sequences, the differences in hly transcript levels observed in all strains harbouring deletions of the hly 5′ UTR, independent of the promoter, indicate that the hly 5′ UTR likely alters gene expression through a post-transcriptional mechanism. Taken together, our results suggest that the hly 5′ UTR may have evolved to maximize expression of LLO, perhaps by increasing the stability of native hly transcripts.

In addition to demonstrating a role for the hly 5′ UTR in regulating LLO expression, we have also shown that the hly 5′ UTR can function to enhance expression of other cis-associated genes. Fusion of the hly 5′ UTR to gfp resulted in a 43-fold increase in GFP fluorescence when the hly 5′ UTR-gfp fusion was transcribed from the HyperSPO1 promoter (Fig. 6C). Furthermore, we observed that the hly RBS was not specifically required for the hly 5′ UTR to enhance heterologous gene expression as an alternate RBS for initiation of translation could be substituted (data not shown). These results demonstrate that the hly 5′ UTR is sufficient to indiscriminately increase expression of downstream cis-associated gene products. Interestingly, the presence of the hly 5′ UTR enhanced expression of heterologous genes to a greater extent than for the native hly gene, even though the presence of the hly 5′ UTR increased transcript levels similarly regardless of the downstream coding sequence (Figs 5 and 6). Taken together, these results may suggest that the hly 5′ UTR enhances gene expression through two mechanisms, acting at a post-transcriptional and translational level. Alternatively, the ability of downstream coding sequences to modulate hly 5′ UTR-mediated enhancement of gene expression may suggest that sequences within the hly gene itself may dampen the full enhancement effect of the hly 5′ UTR. These internal sequences may function to reduce the translational efficiency of hly transcripts harbouring the hly 5′ UTR to prevent production of LLO to levels that may be toxic to host cells during intracellular infection, yet still allow for sufficient LLO to mediate optimal escape from secondary spreading vacuoles. This adaptation would be analogous to the presence of a PEST-like sequence within LLO that modulates intracellular protein levels to prevent toxicity to host cells (Decatur and Portnoy, 2000).

An increasing number of studies have demonstrated the importance of 5′ UTRs in regulating gene expression through diverse mechanisms (Grundy and Henkin, 2004). In this report, we have demonstrated that the hly 5′ UTR enhances expression of LLO during intracellular infection and have shown the ability of the hly 5′ UTR to enhance expression of cis-associated heterologous genes. The precise mechanism by which the hly 5′ UTR is able to mediate this effect remains to be determined, although sequences comprising the entire hly 5′ UTR appear to be required for maximal enhancement of protein expression. Furthermore, the mechanism by which the hly 5′ UTR enhances gene expression appears to be temperature and PrfA-independent, as the hly 5′ UTR was able to enhance expression of reporter genes expressed from the HyperSPO1 promoter independent of both growth temperature and the presence of PrfA (data not shown). Moreover, the ability of the hly 5′ UTR to enhance gene expression does not appear to be a general phenomenon, as the 5′ UTR of L. monocytogenes mpl (Vazquez-Boland et al., 2001) or an artificial 5′ UTR fused to heterologous genes did not result in increased gene expression (data not shown). Lastly, we have taken advantage of the ability of the hly 5′ UTR sequence to indiscriminately increase expression of cis-associated gene products to achieve high-level expression of single-copy, chromosomal gene fusions in L. monocytogenes. Strains harbouring deletions of both hly and prfA could be complemented, independent of PrfA-dependent transcriptional activation, to expression levels greater than those observed for wild-type bacteria by placing hly 5′ UTR gene fusions under transcriptional control of the HyperSPO1 promoter (Fig. 5 and data not shown). The ability to complement virulence gene expression independent of additional regulatory factors will prove beneficial in determining the temporal or expression level requirements of L. monocytogenes virulence determinants during intracellular infection (Dancz et al., 2002; Gründling et al., 2003). Future studies to determine the role of hly 5′ UTR secondary structure in regulating LLO levels during extracellular growth and intracellular infection, as well as assessing the effect of the hly 5′ UTR on transcript stability, will yield further insights into the functional mechanism.

Experimental procedures

Bacterial and eukaryotic cell growth conditions

Bacterial strains used in this study are listed in Table 1. L. monocytogenes cultures were grown at 30°C in BHI broth without agitation. Where indicated, L. monocytogenes cultures were subsequently diluted and grown at 37°C with shaking prior to experimental analysis. Escherichia coli strains were grown in Luria–Bertani (LB) medium at 37°C with shaking. All bacterial stains were stored at −80°C in BHI or LB medium supplemented with 40% glycerol. Antibiotics were used at the following concentrations: carbenicillin, 100 µg ml−1; chloramphenicol, 20 µg ml−1 for selection of pPL2-derived plasmids in E. coli, and 7.5 µg ml−1 or 10 µg ml−1 for selection of integrated pPL2 or pCON1 derivatives in L. monocytogenes respectively. L2 mouse fibroblast cells were grown in RPMI supplemented with 10% fetal bovine serum (HyClone, Logan, UT) and 2 mM glutamine (Mediatech, Herndon, VA). BMM were cultured as described (Portnoy et al., 1988). Eukaryotic cell cultures were maintained at 37°C in a 5% CO2-air atmosphere.

Table 1. Strains used in this study.
StrainDescriptionSource or reference
L. monocytogenes strains
 10403SWild-type strain Bishop and Hinrichs (1987)
 DP-L216110403S Δhly Jones and Portnoy (1994)
 DH-L123110403S with +13 to +113 of the hly 5′ UTR deletedThis study
 DH-L911DP-L2161 with pH-hly PL3 integrated into the tRNAArg locusThis study
 DH-L934DP-L2161 with pH-hlyΔ1-34-PL3 integrated into the tRNAArg locusThis study
 DH-L935DP-L2161 with pH-hlyΔ1-68-PL3 integrated into the tRNAArg locusThis study
 DH-L936DP-L2161 with pH-hlyΔ1-113-PL3 integrated into the tRNAArg locusThis study
 DH-L103910403S with pH-hly gfp-PL3 integrated into the tRNAArg locusThis study
 DH-L111810403S with pH-hlyΔ1-34-gfp-PL3 integrated into the tRNAArg locusThis study
 DH-L111910403S with pH-hlyΔ1-68-gfp-PL3 integrated into the tRNAArg locusThis study
 DH-L104110403S with pH-hlyΔ1-113-gfp-PL3 integrated into the tRNAArg locusThis study
E. coli strains
 DH-E123pCON1 in JM109 Freitag (2000)
 DH-E182XL1-Blue [F′proAB lacIqΔ(lacZ)M15 Tn10]recA1 endA1 gyrA96 thi-1 hsdR17 supE relA1 lacStratagene
 DH-E1232 hlyΔ13-113-pCON1 in XL1-BlueThis study
 DH-E375CLG190 (F′lac pro lacIq) Δ(malF)3 Δ(phoA) PvuII phoRΔ(lac)X74
Δ(ara leu)7697 araD139 galE galK pcnB zad::Tn10 recA; Strr
D. Boyd
 DH-E474SM10 {Fthi-1 thr-1 leuB6 recA tonA21 lacY1 supE44+ C λ
[RP4-2 (Tc::Mu)] Kmrtra+}
Simon et al. (1983)
 DH-E899pHPL3 in XL1-Blue Gründling et al. (2004)
 DH-E897pH-hly PL3 in XL1-BlueThis study
 DH-E937pH-hlyΔ1-34-PL3 in XL1-BlueThis study
 DH-E938pH-hlyΔ1-68-PL3 in CLG190This study
 DH-E939pH-hlyΔ1-113-PL3 in CLG190This study
 DH-E1038pH-hly gfp-PL3 in XL1-BlueThis study
 DH-E1121pH-hlyΔ1-34-gfp-PL3 in XL1-BlueThis study
 DH-E1122pH-hlyΔ1-68-gfp-PL3 in XL1-BlueThis study
 DH-E1075pH-hlyΔ1-113-gfp-PL3 in XL1-BlueThis study
 DH-E846p2RGFP in XL1-BlueThis study

Construction of a L. monocytogenes strain containing a deletion within the hly 5′ UTR

Primers ♯42 and ♯343 (Table 2) and genomic DNA from wild-type L. monocytogenes (10403S) were used to polymerase chain reaction (PCR)-amplify an ∼1 kb product harbouring the 5′ portion of the hly gene. Primers ♯342, ♯398 and 10403S genomic DNA were used in a PCR reaction to generate an ∼700 bp product harbouring a 3′ portion of hly. The 5′ and 3′ PCR products were gel purified using the QIAquick gel extraction kit (QIAGEN, Valencia, CA), and used as templates for a splicing by overlap extension (SOE) PCR reaction (Horton et al., 1989). The flanking primers ♯42 and ♯398 were used to amplify a 1.7 kb PCR product containing a deletion of 101 bp of the hly 5′ UTR sequence (+13 to +113 relative to the native P2 hly transcription initiation site). The SOE PCR product was gel purified, digested with XbaI and KpnI and then ligated with plasmid pCON1 digested with the same restriction enzymes. The resulting plasmid, hlyΔ13–113-pCON1, was confirmed using automated sequencing, introduced into 10403S by electroporation and allelic exchange was performed as previously described (Camilli et al., 1993) to generate strain DH-L1231.

Table 2. Oligonucleotides used in this study.
NumberSequenceSitea
  • a

    . The indicated restriction endonuclease site is underlined within the oligonucleotide sequence.

 42CCTCTAGACGGGGAAGTCCATGATTAGTATGCCXbaI
 68GCAGATGCATCCTTTGCTTCAGTTTG
133CGCAGCAAATGCTGTTACCGCAATCCCAGCTGTAGCCGCG
168AGATACCGGCCGATAAAGCAAGCATATAATATTGCGTTEagI
169AGATACCGGCCGAGAAGCGAATTTCGCCAATATTATAATTATEagI
205AGATACCGGCCGAGAGAGGGGTGGCAAACGGTATTEagI
221AAGTCGACTTATTTGTATAGTTCATCCATGCCATGSalI
235AGATACCGGCCGGTAGAAGGAGAGTGAAACCCEagI
245GTGCGTCGTAAATAAATCTTATACAA
250CGCGGATCCATTTACGAAGAGTGCAAAACAAGCBamHI
341ACAACTCCAGTGAAAAGTTCTTCTCCTTTACTCAT
342ATAAAGCTATAAAGCAAGCAGTAGAAGGAGAGTGAAACCC
343GGGTTTCACTCTCCTTCTACTGCTTGCTTTATAGCTTTAT
344GTAGAAGGAGAGTGAAACCCATGAGTAAAGGAGAAGAACT
345AGTTCTTCTCCTTTACTCATGGGTTTCACTCTCCTTCTAC
398GGGGTACCCTTTAAATGCTGTACCAAATTTCGCKpnI

Haemolytic-activity assays

Fourteen to 16 h cultures of L. monocytogenes strains grown in BHI were diluted 1:10 in BHI broth and grown for 5 h at 37°C. The optical density at 600 nm (OD600) was determined. One millilitre aliquots were removed and the haemolytic activity present in culture supernatants was determined as described (Dancz et al., 2002). Haemolytic units were defined as the reciprocal of the dilution of culture supernatant that yielded 50% lysis of sheep red blood cells.

Western blot analysis of bacteria grown in BHI broth

Fourteen to 16 h cultures of L. monocytogenes strains grown in BHI broth were diluted 1:10 in BHI or BHIC broth and grown for 5 h at 37°C. The OD600 of cultures was determined. For cell pellet fractions (CP), 1 ml aliquots were removed and centrifuged. The supernatant was removed for haemolytic activity determination as described, while the bacterial cell pellets were incubated in TE-lysozyme (2 mg ml−1 lysozyme) buffer at 37°C in a total volume of 100 µl. Following digestion for 2 h, 100 µl of 2× final sample buffer (FSB) was added to CP samples. For analysis of supernatant fractions (SN), 1.2 ml aliquots were centrifuged and 1 ml of the resulting supernatant was incubated with 100 µl of 100% trichloroacetic acid (TCA) on ice. Proteins were TCA-precipitated for 1 h and then centrifuged at 13 000 g for 10 min. Protein samples were resuspended in 200 µl of 1× FSB containing 0.1 N NaOH. Samples in FSB were boiled for 5 min at 95°C, centrifuged for 5 min at 13 000 g and loaded onto denaturing 10% or 12% polyacrylamide gels. Protein samples from a culture volume equivalent to 100 µl (CP) or 50 µl (SN) of an OD600 = 1.5 were analysed for LLO. Protein samples from a culture volume equivalent to 100 µl of an OD600 = 1.0 were analysed for GFP. Western blotting was performed using rabbit polyclonal anti-LLO (Dancz et al., 2002) or Living Colors A.v. peptide anti-GFP antibodies (Clontech, Palo Alto, CA). Densitometry analysis was performed using ImageQuant TL software (Amersham Biosciences, Piscataway, NJ).

RNA isolation

Fourteen to 16 h cultures of L. monocytogenes grown in BHI broth were diluted 1:10 in BHI broth and grown for 5 h at 37°C, while strains used for gfp transcript analysis were grown for 3 h at 37°C. Total RNA was isolated from 6 ml aliquots of cultures using the FastRNA ProBlue kit (Qbiogene, Carlsbad, CA) according to the manufacturer's instructions.

Primer extension analysis

Oligonucleotide primers ♯68, ♯245, ♯341 and ♯133 were used for primer extension analysis of hly, plcA, gfp and iap transcripts respectively. Primer extension was performed essentially as described by (Gründling et al., 2004) with the following modifications. T4 polynucleotide kinase was heat-inactivated by incubating the kinase reaction at 70°C for 10 min; 7.5 µg total RNA was used in each extension reaction, with the exception of Fig. 3C, where 15 µg total RNA was used for samples isolated in BHI broth, and 3 µg total RNA was used for samples isolated from bacteria grown intracellularly. Extended transcripts were quantified by phosphorimager analysis (Molecular Imager, Bio-Rad, Hercules, CA, and Typhoon Imager, Amersham Biosciences, Piscataway, NJ).

Immunoprecipitation of secreted LLO during infection of J774 cells

Approximately 2.2 × 106 J774 cells were seeded on tissue-culture treated 60 mm dishes and incubated for 12–15 h. Monolayers of J774 cells were then infected in duplicate with 2.5 µl (∼5 × 106 bacteria) of a PBS-washed 14–16 h culture of L. monocytogenes in 6 ml of DMEM. After 1 h of infection, the monolayer was washed once with PBS, pH 7.1, and 6 ml of DMEM with 5 µg ml−1 gentamicin was added. At 5.5 h after infection, cells were washed once with PBS, pH 7.1, and starved in methionine-free DME (Met-DME) containing 7.5% FBS, 5 µg ml−1 gentamicin, 30 µg ml−1 anisomycin, 100 µg ml−1 cycloheximide, and 50 µM LLnL. After 30 min, one monolayer was pulse-labelled for 1 h with 200 µCi of 35S-methionine (EasyTag protein labelling mix, Perkin Elmer, Boston, MA) in 450 µl of Met-DME, while the second monolayer was mock-labelled with cold methionine at 8.5 µM in 450 µl of Met-DME for subsequent colony-forming unit (cfu) determination. Monolayers were then washed, and lysed in 0.5 ml of ice-cold RIPA buffer [150 mM NaCl, 50 mM Tris-HCl (pH 8.0), 1% NP-40, 0.1% SDS] containing Complete protease inhibitor mixture (Roche, Indianapolis, IN). Nuclear insoluble material was removed by pulsing host cell lysates for 2 s at 13 000 g, and bacteria were isolated from the resulting supernatant by centrifugation for 7 min at 13 000 g. For non-radioactive samples, the resulting pellet was resuspended in 40 µl of PBS and cfu were determined. For radioactive samples, the supernatant was transferred to 250 µl of a 50% Protein A-Sepharose bead slurry preconjugated to monoclonal anti-LLO antibody B3-19. LLO was immunoprecipitated as previously described (Moors et al., 1999b). Immunoprecipitated samples were boiled and resolved on an 8% polyacrylamide gel by SDS-PAGE. The resulting gel was exposed to a phosphorimager screen for 7 days and quantified by phosphorimager analysis (Typhoon Imager, Amersham Biosciences, Piscataway, NJ).

Western blot analysis and RNA isolation of bacteria grown in J774 cells

Approximately 5.0 × 106 J774 cells were seeded on tissue-culture treated 100 mm Petri dishes 12–15 h prior to infection. Monolayers of J774 cells were then infected in duplicate with 5.0 µl (∼1 × 107 bacteria) of a PBS-washed 14–16 h culture of L. monocytogenes in 10 ml of DMEM. After 1 h of infection, the monolayers were washed once with PBS, pH 7.1, and 10 ml of DMEM with 5 µg ml−1 gentamicin was added. At 7 h post infection, 1 ml of ABT buffer (60 mM K2HPO4, 40 mM KH2PO4, 100 mM NaCl, 0.1% Triton X-100, pH 7.0) was added to lyse J774 cells. Nuclear insoluble material was removed by pulsing the host cell lysates for 2 s at 13 000 g, and the bacteria were pelleted from the resulting supernatant by centrifuging for 7 min at 13 000 g. For one monolayer, the bacterial cell pellet was processed for RNA isolation using the FastRNA ProBlue kit (Qbiogene, Carlsbad, CA) according to the manufacturer's instructions. For the second monolayer, the bacterial cell pellet (CP) was resuspended in TE-lysozyme (2 mg ml−1 lysozyme) buffer at 37°C in a total volume of 40 µl, and 2.5 µl was removed for cfu determination. Following digestion for 30 min at 37°C, 40 µl of 2× FSB was added to CP samples. Samples in FSB were boiled for 5 min at 95°C, centrifuged for 5 min at 13 000 g, and a sample equivalent to 2 × 108 cfu was loaded onto a 10% denaturing polyacrylamide gel. Western blot analysis was performed using monoclonal anti-LLO antibody (B3-19). Densitometry analysis was performed using ImageQuant TL software (Amersham Biosciences, Piscataway, NJ).

Intracellular growth of L. monocytogenes in murine BMM

Bone marrow from the femurs of adult BALB/c mice was isolated and BMM cultured as described (Portnoy et al., 1988). Approximately 6.6 × 105 bacteria were used to infect 2 × 106 BMM cells seeded 18 h prior onto circular 12 mm glass coverslips placed in a 60 mm Petri dish. Infection of BMM was performed in DMEM (Mediatech, Herndon, VA) supplemented with 7.5% fetal bovine serum (HyClone, Logan, UT) and 2 mM glutamine (Mediatech, Herndon, VA). Thirty minutes after addition of bacteria, BMM were washed with PBS, and DMEM containing 10 µg ml−1 gentamicin was added. Three coverslips were removed at appropriate time intervals and separately placed in 5 ml of sterile dH2O in a 15 ml conical tube. Conical tubes were vortexed for 15 s to lyse BMM and dilutions of lysates were plated on LB-agar to determine the number of intracellular bacteria.

Plaque formation assays in L2 fibroblasts and determination of LD50 values

Assays of plaque formation within L2 cell monolayers were performed as described (Dancz et al., 2002). The average diameter of 15 plaques/sample for three independent experiments was determined. LD50 values were determined as described (Barry et al., 1992).

Construction of HyperSPO1 promoter-controlled hly strains

To construct DH-L911, harbouring the full-length hly 5′ UTR sequence fused to hly, primer pair ♯168 and ♯250 was used to PCR-amplify the hly gene from 10403S genomic DNA. The resulting PCR product was digested with EagI and BamHI and ligated with the pPL2-derived vector pHPL3 (Lauer et al., 2002; Gründling et al., 2004) digested with the same restriction enzymes to yield plasmid pH-hly PL3. To construct plasmid vectors lacking portions of the hly 5′ UTR sequence, each of the following forward primers, ♯169, ♯205 or ♯235, was used separately with reverse primer ♯250 and pH-hly PL3 as the template to PCR-amplify a DNA fragment containing the entire hly coding sequence with varying lengths of upstream DNA sequences. The resulting PCR products were digested with EagI and BamHI and ligated with pHPL3 digested with the same restriction enzymes. The sequence of all DNA inserts was confirmed by automated sequencing. The resulting plasmid from ligation of the PCR product generated using primer pair ♯169 and ♯250, pH-hlyΔ1−34-PL3, was transformed into E. coli strain XLI-Blue. Triparental mating (Lauer et al., 2002) was used to transfer pH-hly PL3 and pH-hlyΔ1−34-PL3 into DP-L2161. The resulting plasmid from ligation of the PCR product generated using primer pair ♯205 and ♯250, pH-hlyΔ1−68-PL3, was transformed into E. coli strain CLG190. The resulting plasmid from ligation of the PCR product generated using primer pair ♯235 and ♯250, pH-hlyΔ1−113-PL3, was also transformed into CLG190. pH-hlyΔ1−68-PL3 and pH-hlyΔ1−113-PL3 were isolated from CLG190 strains using the Qiagen Midiprep kit (Qiagen, Valencia, CA), dialysed, and transformed into electrocompetent DP-L2161 by electroporation. Integration of plasmid vectors onto the chromosome was selected as described (Lauer et al., 2002).

Construction of HyperSPO1 promoter-controlled gfp strains

Fusion of hly 5′ UTR sequences to gfpmut2 was performed using SOE PCR. Primers ♯168 and ♯345 were used with genomic DNA from 10403S to generate the 5′ PCR fragment, while primers ♯344 and ♯221 were used to amplify a 3′ PCR fragment from plasmid p2RGFP harbouring the gfpmut2 gene. The resulting PCR products were gel purified and used as templates in an SOE PCR reaction using the flanking primers ♯168 and ♯221. The final SOE PCR product containing the entire hly 5′ UTR sequence fused to the initiating codon of gfpmut2 was gel purified, digested with EagI and SalI, and ligated with the pHPL3 vector digested with the same restrictions enzymes. The resulting plasmid, pH-hly gfp-PL3, was used as a template in PCR reactions to generate gfp fusions to hly 5′ UTRs of decreasing length. Forward primers ♯169, ♯205, or ♯235 were used separately in PCR reactions with reverse primer ♯221 and pH-hly gfp-PL3. PCR products were gel purified, digested with EagI and SalI, and ligated to pHPL3 digested with the same restriction enzymes to generate plasmids pH-hlyΔ1−34-gfp-PL3, pH-hlyΔ1−68-gfp-PL3 and pH-hlyΔ1−113-gfp-PL3. pH-gfp-PL3 plasmid derivatives were verified by automated sequencing and transformed into electrocompetent 10403S by electroporation.

Analysis of GFP expression

Western blot analysis of GFP levels was performed as described, with the exception that cultures were grown for 3 h instead of 5 h. The relative fluorescence values of cultures used for Western blot analysis were also determined. One millilitre aliquots were centrifuged and bacterial pellets resuspended in 1 ml of PBS, pH 7.1. Triplicate aliquots of 200 µl were transferred to a black, polystyrene, flat bottom 96 well plate (Corning, Corning, NY) and GFP fluorescence activity determined using a SpectraMAX GeminiXS fluorimeter (Molecular Devices, Sunnyvale, CA). An excitation wavelength of 488 nm, an emission range of 525 nm, and an auto cut-off of 515 nm were used to minimize interfilter effects. Relative fluorescence of samples was determined by measuring the background fluorescence of 10403S and subtracting this value from the fluorescence units measured for 10403S strains expressing GFP to yield relative fluorescence units (RFU). RFU were normalized against OD600 to give RFU/OD600.

Acknowledgements

We thank H.G. Archie Bouwer for performing the virulence studies in mice. We would also like to thank Hervé Agaisse for kindly providing the protocol for Northern blot analysis, Bryan Krantz for advice with fluorimeter measurements, Daniel Portnoy for monoclonal anti-LLO antibody B3-19, and Pamela Schnupf for kindly providing the protocol for LLO immunoprecipitations. This work was supported by US Public Health Service Grant AI53669 from the National Institutes of Health (D.E.H.). A.S. is a recipient of a Howard Hughes predoctoral fellowship award.

Ancillary