INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Listeria monocytogenes is an important bacterial pathogen that causes severe infections in many species of animals, including livestock and humans. The manifestations of listeriosis include encephalomeningitis, septicemia, and abortion in pregnant women (40). L. monocytogenes is a facultative intracellular organism that is able to grow and survive under a wide variety of conditions, including in soil, plants, water, and mammalian tissues, and is widely distributed in the environment (reviewed in reference 16). Thus, L. monocytogenes must be able to sense these different environments and respond to them by regulating the proper repertoire of genes in a manner that ensures the optimal growth of the bacterium.

L. monocytogenes genes have been identified that are required for growth within mammalian cells. Several of these genes are located on the L. monocytogenes chromosome at a single locus. This locus encodes two phospholipases C, PlcA and PlcB, that together enable Listeria to escape both from the primary phagosome after uptake and from the double membrane vesicles formed as a result of cell-to-cell spread (7, 37, 41, 44). Mpl, a metalloprotease that processes PlcB (10, 30), and ActA, an actin polymerization protein necessary for movement of Listeria within the cytoplasm of the host cell and into adjacent cells, are also encoded within this locus (11, 25). Also present is hly, which encodes listeriolysin O (LLO), a hemolysin required for the escape of Listeria from both the primary phagosome and the double membrane vesicles formed as a result of cell-to-cell spread (3, 20, 35). The transcriptional regulator of these genes, PrfA, also at this locus, is required for the induction of this important virulence gene cluster during the invasion process (8, 26, 29). In addition, PrfA regulates several other virulence genes, including inlA, inlB, and irpA (inlC), that are located elsewhere on the Listeria chromosome (12, 13, 15, 32).

The importance of these genes in the virulence of Listeria has been examined in the mouse model of listeriosis. L. monocytogenes harboring mutations in actA (5) or hly (2) exhibits a 1,000- and 10,000-fold increase in 50% lethal dose in the mouse model of infection, while plcA or plcB mutants show a 2- and 20-fold reduction in virulence, respectively (7, 41). Reported 50% lethal doses for irpA (inlC) mutants indicate a 2- to 50-fold reduction in virulence in mice (12, 15). Many of these virulence genes are preferentially expressed when Listeria is within the host cell (6, 18, 33). We initiated a fluorescence-activated cell sorting (FACS) screen in an attempt to identify additional L. monocytogenes genes that are induced in vivo and may also be required for the pathogenesis of this important bacterial pathogen.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids Bacterial strains, plasmids, and primers used in this study are described in Tables 1 and 2.

Bacterial growth and tissue culture. Escherichia coli strains were grown in Luria broth (39). L. monocytogenes strains were grown in brain heart infusion (BHI) broth (Difco) or tryptic soy broth (Difco). The following antibiotics were used at the indicated concentrations (µg/ml): ampicillin, 100; chloramphenicol, 10; and streptomycin, 50. The murine macrophage-like cell line J774 was grown in RPMI 1640 (Gibco) containing 10% fetal calf serum and supplemented with 2 mM L-glutamine, 5 mm HEPES buffer, 50 µm 2--mercaptoethanol, 100 U (each) of penicillin and streptomycin, and 50 µg of gentamicin sulfate per ml (RP10). Cell lines were maintained at 37°C in a humidified atmosphere of 7% CO₂.

Plasmid and library construction. The promoterless green fluorescent protein (GFP) shuttle plasmid, pAMGFP, was constructed by restriction digestion of plasmid pGFPmut3 (9) with PstI. The ends were made blunt with T4 DNA polymerase, and the plasmid was digested with BamHI. After agarose gel purification, the GFP-encoding DNA fragment was ligated to BamHI-EcoRV-digested plasmid pAM401 (46), forming plasmid pAMGFP.

Electroporation of L. monocytogenes Listeria was made electrocompetent using a method similar to that described by Michel et al. (31). Briefly, an overnight culture of L. monocytogenes was diluted 1/100 into 100 ml of BHI medium and grown to an optical density at 600 nm (OD₆₀₀) of 0.5 to 0.8. Cells were centrifuged at 4°C for 15 min at 5,000 × g, and the cell pellet was suspended in 30 ml of ice-cold electroporation buffer (816 mM sucrose, 3 mM MgCl₂). The centrifugation and suspension steps were repeated two more times, and the final pellet was suspended in 4 ml of electroporation buffer. One hundred microliters of cells was aliquoted into cold Eppendorf tubes and stored at 80°C until needed.

Flow cytometry and FACS. To enrich cultures for Listeria expressing GFP within macrophages, 3 × 10⁶ J774 cells were plated in six T25 flasks in RP10 medium lacking antibiotics, and the plates were incubated overnight. Cultures of DP-L1942 transformed with either the Listeria GFP library or the promoterless pAMGFP plasmid alone were grown to exponential phase (optical density at 600 nm [OD₆₀₀]  0.1) in BHI containing chloramphenicol and streptomycin. The cultures were centrifuged and suspended in RP10 containing chloramphenicol and streptomycin and were used to infect the J774 cells at an approximate multiplicity of infection (MOI) of 1. After a 2.5-h infection, the medium was removed and RP10 containing gentamicin, chloramphenicol, and streptomycin was added for 1.5 h. The infected macrophages were collected, and the 1% of cells expressing the highest levels of GFP were isolated by FACS using a Coulter EPICS 753 cell sorter at the University of Iowa Flow Cytometry Facility. Approximately 3,000 events were collected, the J774 cells were centrifuged, and the pellet was lysed in 500 µl of 1% Triton X-100 to release the bacteria. After 10 min at room temperature, 500 µl of RP10 was added and the bacteria were plated on BHI agar with chloramphenicol and streptomycin. Approximately 14,000 colonies were pooled (called library A) and frozen in aliquots at 80°C in 30% glycerol. To separate bacteria carrying inducible GFP from those with constitutively expressed GFP, a 200-µl aliquot of library A was thawed and grown in RP10 containing chloramphenicol and streptomycin for 4 h. The culture was sorted by FACS, and the 11% of bacteria expressing the lowest levels of GFP was collected. This sorted population was used to infect J774 cells as described above, and the 0.6% of fluorescent macrophages expressing the highest levels of GFP was collected. The macrophages were lysed and the bacteria were collected as described above. This population of bacteria was termed library B. Individual colonies from library B were tested for in vivo expression of GFP as described below.

In vitro infections J774 cells were plated at 2 × 10⁵/ml in RP10 without antibiotics in 24-well tissue culture plates and grown overnight. Single colonies from library B or the control strain DP-L1942(pAMGFP) were grown overnight and subcultured 1:50 the following day into BHI medium containing chloramphenicol and streptomycin. After 1 to 2 h of growth at 37°C, 500 µl of culture was used to infect J774 cells (MOI, ~200). Following a 1.5-h incubation at 37°C the J774 cells were washed one time with phosphate-buffered saline, and 1 ml of RP10 containing gentamicin (50 µg/ml) was added for another 2.5 h. The infected J774 cells were lifted using a cell scraper and transferred to tubes suitable for flow cytometry analysis.

DNA sequence analysis Plasmids were isolated from Listeria clones using Qiagen plasmid mini prep kits according to the manufacturer's instructions except that 15 µl of lysozyme (50 mg/ml) was added to 250 µl of cell suspension buffer, and cells were incubated 2 to 24 h before proceeding. The DNA sequences of in vivo-induced Listeria promoters were determined using a primer specific for the gfp gene (GFP-3') in plasmid pAMGFP or primers designed from the determined DNA sequence of the individual clones. Sequencing was done at the University of Iowa DNA Core Facility. Homology searches were performed using the databases at the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov).

Mutant construction To construct an L. monocytogenes uracil DNA glycosylase (ung) mutant, a plasmid clone (pG1-E1) containing the entire ung gene was identified in a previously constructed L. monocytogenes 10403s library (27) using PCR with primers specific for ung: 117-5'2 and 117-3'. The DNA sequence of the region upstream and downstream of ung and that of the entire ung gene were determined. Primers specific for plasmid pAMGFP (GFP-5') and 157 bp downstream of the predicted ATG start codon of ung (117-BamHI) were constructed and used to PCR amplify the 5' region of ung using plasmid DNA from clone 117 as a template. This PCR product was digested with EcoRI and BamHI restriction enzymes and gel purified. The 3' region of ung was PCR amplified using primers 117-5'3 and 117-PstI with plasmid pG1-E1 as a template. The PCR product was digested with PstI and Sau3AI restriction enzymes and gel purified. These two PCR products were used in a triple ligation to plasmid pKSV-7 that had been digested with EcoRI and PstI and gel purified. The resulting clone was designated p117/KSV and contains a 68-amino-acid in-frame deletion (amino acids 53 to 121) of the predicted 228-amino-acid UDG protein. Plasmid p117/KSV was used to transform electrocompetent L. monocytogenes 10403s, and transformants were selected on BHI medium containing chloramphenicol and streptomycin at 30°C. Integration of this temperature-sensitive plasmid into the chromosome and resolution of the plasmid were performed as previously described (7). The resulting chloramphenicol-sensitive colonies were screened by PCR using primers 117-5'BamHI and 117-PstI to identify those which contained the ung deletion in the proper chromosomal location.

Mouse virulence studies Bacteria were grown overnight in BHI medium, subcultured 1:1,000 in BHI medium the following day, and grown to an OD₆₀₀ of ~0.1. Bacteria were centrifuged and suspended at 10⁵ CFU/ml in pyrogen-free saline. Groups of BALB/c mice (National Cancer Institute) were infected intravenously with 4 × 10⁴ L. monocytogenes organisms in a 0.2-ml volume. Three days postchallenge the number of CFU in the spleen and liver of the infected animals was determined by homogenizing tissues in 0.2% IGEPAL (Sigma) and plating dilutions of the suspension on BHI agar containing streptomycin.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of a gram-positive shuttle plasmid encoding a promoterless GFP. A plasmid encoding a promoterless GFP gene was constructed by inserting DNA encoding the Aequorea victoria GFP from plasmid pGFPmut3 into the gram-positive shuttle plasmid pAM401 (46) (Fig. 1). pGFPmut3 encodes a mutated form of GFP that exhibits maximal excitation and emission wavelengths at 501 and 511 nm, respectively, and is useful for applications utilizing flow cytometry (9). To test whether an in vivo-induced Listeria promoter could drive expression of GFP from this plasmid (pAMGFP), the hly promoter was cloned into the BamHI site of pAMGFP. Expression of hly is known to be induced immediately after L. monocytogenes enters the host cell (6, 33). This plasmid, called pLLOGFP, was introduced into the attenuated actA mutant DP-L1942 (5). The actA mutant was chosen for these studies because we reasoned that if L. monocytogenes could not spread from the initially infected cells into adjacent cells, the bacteria would accumulate, and the host cells containing the GFP-expressing bacteria would be more fluorescent due to increased numbers of bacteria. DP-L1942(pLLOGFP) was used to infect J774 cells, a macrophage-like cell line. As shown in Fig. 2, J774 cells infected with DP-L1942(pLLOGFP) exhibited higher fluorescence than the same cells infected with DP-L1942(pAMGFP). Hence, the level of expression of GFP from plasmid pAMGFP containing an in vivo-induced L. monocytogenes promoter is sufficient to distinguish fluorescent from nonfluorescent infected macrophages.

View larger version (17K): [in this window] [in a new window]   FIG. 1.   Construction of plasmid pAMGFP. A promoterless gfp mut3 gene encoding the A. victoria GFP was removed from plasmid pGFPmut3 (9) and ligated into the BamHI-EcoRV sites of plasmid pAM401 (46). Sau3AI fragments of the Listeria 10403s chromosome were ligated into the BamHI site of pAMGFP to construct the library used for FACS for in vivo-induced promoters.

View larger version (25K): [in this window] [in a new window]   FIG. 2.   Expression of pLLOGFP by L. monocytogenes increases the fluorescence of infected J774 cells. L. monocytogenes DP-L1942 containing plasmid pAMGFP or plasmid pLLOGFP was used to infect J774 cells at an MOI of 10. After 4 h, the infected J774 cells were analyzed by flow cytometry.

FACS for in vivo-induced L. monocytogenes genes. To identify in vivo-induced genes, a library of random, Sau3AI-digested L. monocytogenes DNA fragments was cloned into pAMGFP, and the library was introduced into the actA mutant DP-L1942. Approximately 5,600 different transformants were pooled. With an average DNA fragment size of 600 bp and a probability of 95%, this library was calculated to represent about one-fourth of the Listeria chromosome. A technique similar to that developed by Valdivia and Falkow for Salmonella (43) was employed to identify in vivo-induced L. monocytogenes genes by FACS. The first sort was performed on J774 cells that had been infected for 4 h with the pooled L. monocytogenes containing the GFP library (Fig. 3A). The 1% of macrophages exhibiting the highest fluorescence was collected and lysed to isolate the infecting bacteria. A second sort was done on the released bacteria after growth in tissue culture medium alone (Fig. 3B), and the 11% of bacteria with the lowest fluorescence was collected. This step was performed to remove those bacteria from the population that constitutively expressed GFP. J774 cells were infected with this population of bacteria exhibiting low extracellular fluorescence, and a final sort was done to collect the 0.6% of fluorescent macrophages with the highest GFP expression (Fig. 3C). Bacteria released from this population of macrophages should be enriched for those containing plasmids carrying in vivo-induced promoters.

View larger version (16K): [in this window] [in a new window]   FIG. 3.   FACS for in vivo-induced promoters. (A) Fluorescence of J774 cells infected with DP-L1942 (actA) containing the pAMGFP library. The solid line indicates that the sorting gate was set to collect the 1% of cells expressing the highest fluorescence. (B) Fluorescence of the population of bacteria released from the macrophages sorted in the step depicted in panel A. The 11% of bacteria expressing the lowest fluorescence was collected. (C) Fluorescence of J774 cells infected with bacteria from the step depicted in panel B. The line indicates that the sorting gate was set to collect the 0.6% of cells expressing the highest GFP.

Identification of sorted L. monocytogenes clones encoding potential in vivo-induced genes. In order to identify the individual Listeria clones that contain plasmids carrying in vivo-induced genes, isolated clones were grown in BHI medium and used to infect J774 macrophages. Those clones that caused increased fluorescence of J774 cells after a 4-h infection but were not fluorescent in BHI medium, as compared to Listeria containing a promoterless GFP plasmid, were identified as potential candidates (Fig. 4). Some selected clones had GFP expression profiles that showed no induction in BHI medium and substantial induction within J774 cells (clones 117, 136, and 159). Other clones had a partially induced population when the bacteria were grown in BHI medium, while both populations showed further induction within J774 cells (clones 44, 53, 98, 104, 107, and 154). The remaining clones showed either no induction of GFP in BHI medium or within J774 cells or constitutive expression of GFP (clone 8) under either growth condition. Of 167 isolates tested, 42 clones were selected that showed low levels of GFP expression in BHI medium and higher levels of GFP expression within J774 cells, and the DNA sequence of the region immediately upstream of GFP was determined. These clones and their closest homologues are listed in Table 3. Seventeen of the plasmids encoded a 5' sequence identical to that encoded by the in vivo-induced L. monocytogenes actA gene, which confirms the validity of the cell-sorting technique as a means of identifying in vivo-induced promoter fragments.

View larger version (42K): [in this window] [in a new window]   FIG. 4.   Flow cytometry analysis of individual clones expressing GFP. Overnight cultures of isolated clones were subcultured into fresh BHI medium (containing streptomycin and chloramphenicol) and grown to an OD600 of approximately 0.1. These cultures were used to infect J774 cells at an MOI of approximately 200 and incubated for 4 h at 37°C. The inoculating cultures (BHI) and the infected J774 cells were analyzed by flow cytometry, and their fluorescence is shown (in black) relative to DP-L1942 containing the promoterless pAMGFP plasmid (in grey). The numbers represent the numbers of the individual clones. The fluorescence of bacteria containing the actA promoter in the GFPfusion plasmid that was isolated during the enrichment and the fluorescence of bacteria containing plasmid pLLOGFP are shown for reference. Clone 8 is a representative clone that expressed GFP both in BHI medium and within J774 cells.

Construction of Listeria chromosomal mutants and virulence studies. We wished to determine whether some of the L. monocytogenes in vivo-induced genes identified in our screen were required for virulence in mice. In-frame deletion mutations of the yhdP (clone 104) and ung (clone 117) genes were constructed and introduced into the chromosome of wild-type 10403s L. monocytogenes. The yhdP gene was chosen because of its intriguing homology to the Rickettsia typhi TlyC hemolysin (36) and hemolysin-like homologues (Fig. 5). In E. coli, the function of uracil DNA glycosylase (UDG), the product of the ung gene, is to remove uracil residues that have been misincorporated into DNA as a result of the deamination of cytosine to uracil (28). DNA damage could potentially occur within the environment of the host cell phagosome through the action of nitric oxide and the products of the respiratory burst (45). Growth curves determined for the Listeria yhdP and ung mutants indicated that they had no apparent growth defects in rich broth culture (not shown). The mutants were inoculated into BALB/c mice intravenously, and CFU assays were performed on the livers and spleens after 3 days of infection. Figure 6 demonstrates that a nearly sevenfold decrease in CFU in livers and a threefold decrease in CFU in spleens were measured when mice were infected with the L. monocytogenes yhdP mutant, as compared to those infected with wild-type L. monocytogenes. Deletion of the L. monocytogenes ung gene did not appear to have a demonstrable effect on virulence in the mouse model of L. monocytogenes infection (data not shown).

View larger version (79K): [in this window] [in a new window]   FIG. 5.   ClustalW 1.8 alignment (22) and box shade analysis of the B. subtilis YhdP (accession no. 007585), R. typhi TlyC (accession no. AAC62436), and L. monocytogenes YhdP (clone 104) protein sequences. The boxed gray shading indicates identical protein residues and the boxed light shading indicates similar protein residues.

View larger version (24K): [in this window] [in a new window]   FIG. 6.   An L. monocytogenes yhdP mutant is attenuated in the mouse model of infection. L. monocytogenes 10403s and L. monocytogenes 10403s yhdP were used to infect groups of BALB/c mice intravenously with 2 × 104 to 4 × 104 L. monocytogenes organisms. Data represent the mean ± standard error of the mean (error bars) of CFU per organ from organs of nine mice pooled from three separate experiments. Student's t test was used in statistical analysis. The numbers of CFU per liver (P < 0.005) and per spleen (P < 0.025) in yhdP mutant-infected mice were reduced compared to those in L. monocytogenes 10403s-infected mice.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Using a FACS technique similar to that developed by Valdivia and Falkow (43), we have identified several L. monocytogenes genes that appear to be induced when the bacteria are in the intracellular environment of a macrophage. Several clones that were isolated contained DNA from the L. monocytogenes actA gene, which encodes one of the proteins necessary for the nucleation of host cell actin and motility of Listeria within the host cell. The actA gene is induced about 200-fold in vivo (33). This result provides strong support for the validity of this method for identifying Listeria genes that are expressed within macrophages. One advantage of using promoter trap systems as described here is that plasmid reporter gene fusions are utilized, and these do not depend upon inactivation of the chromosomal locus, which could lead to the selective isolation of only certain genes or polar effects on downstream genes. A potential limitation of a multicopy-plasmid-based system is, however, that it could conceivably limit the sensitivity of the system by titrating factors required for gene induction.

For the FACS technique to be successful, it is likely that high levels of expression of GFP from the plasmids are necessary to achieve efficient separation of fluorescent from nonfluorescent bacteria. In some systems, the conditions for enrichment of in vivo-induced genes may need to be altered from the original method described (43). For example, a screen for Mycobacterium marinum in vivo-induced genes was described in which a step to isolate bacterium-containing vesicles from infected macrophages was included (1). This allowed the isolation of weak promoter-GFP fusions and contributed to the success of the enrichment. In this study we utilized a Listeria actA mutant to allow the bacteria to accumulate within the host cell and thus increase the fluorescence of macrophages infected with GFP-expressing Listeria. Another change from the original protocol was that individual clones were screened for GFP expression while still within the macrophage. We found this necessary because it was difficult to differentiate the fluorescent bacteria from the lysed macrophage cell debris. Because the cell debris was similar in size but not fluorescent, this resulted in a measured decrease in fluorescence of the bacterial population after passage through macrophages. This decrease was seen even with bacteria containing the actA::GFP fusion plasmid. Although these differences in the protocol may have limited our selection of positive clones, they most likely contributed to the success of the enrichment.

This study identified several L. monocytogenes genes that are induced within the host cell. One of these was the L. monocytogenes bvrA gene, which encodes an antiterminator for bvr (clone 53) (Table 3). The bvr locus was previously shown to be necessary for repression of virulence genes (plcB) when Listeria is grown in the presence of -glucosides (4). It is postulated that -glucosides are present in plant-rich soil, and Listeria would most likely repress virulence genes in this environment. In response to appropriate signals, the BvrA protein is activated to allow transcription of the downstream bvrBC genes, which encode a -glucoside-specific enzyme II permease and a putative ADP-ribosylglycohydrolase. We do not understand the apparent contradiction between the fact that expression of the bvr locus is required for repression of plcB and that an activator of bvr was identified in our screen as an in vivo-induced gene. However, it is interesting that an E. coli i484 in vivo expression technology screen identified bglF as a gene that is induced within the mouse liver (23). bglF encodes the -glucoside-specific phosphotransferase transport system (PTS) protein for E. coli. Further studies are required to understand the contribution of the bvr locus to the intracellular growth of L. monocytogenes.

A screen for Listeria in vivo-induced genes performed by Gahan and Hill (19) identified a gene homologous to that encoding the E. coli cellobiose transporter CelB, a PTS enzyme II. We also identified another phosphotransferase enzyme II component homologue for the E. coli mannose permease system (clone 159) as an in vivo-induced gene. Although it has been hypothesized that the high-energy PTS systems may be downregulated due to the unavailability of their transported carbon sources in vivo (38), the regulation of the carbon utilization systems is most likely complex. One clone that was isolated at least 10 times showed homology to the Bacillus subtilis xylose repressor protein. This clone (clone 44) showed various degrees of GFP expression in BHI medium alone but high GFP expression levels within J774 macrophages. A screen for virulence factors in Streptococcus pneumoniae identified a gene showing homology to a Borrelia burgdorferi xylose operon regulatory protein (34). The function of the XylR-like protein in Listeria is not known, but the isolation of the clone containing xylR-like sequences may be another indication of the importance Listeria places on regulating genes involved in carbohydrate utilization during invasion of mammalian cells. Carbon sources available within the environment of the host cell may signal for the increased expression of certain permeases in order for Listeria to utilize those energy sources.

Another gene, one encoding a putative UDG, was also identified in this screen. Because the regulation of the ung fusion appeared very similar to that of the actA (i.e., no expression in BHI and significant induction within macrophages) (Fig. 4) and because of the role of UDG in DNA repair (28), we constructed an L. monocytogenes ung mutant. However, this ung mutant was not attenuated for virulence in mice. Thus, not all in vivo-induced genes play a required role in pathogenesis.

Clone 104 contained DNA encoding a protein most similar to a hypothetical B. subtilis hemolysin-like protein, YhdP. A motif search of the amino acid sequence of this protein using the Prosite analogue revealed probable integral membrane structures (90% similarity to ABC-2-type transport system integral membrane proteins). An R. typhi hemolysin (TlyC), capable of lysing sheep erythrocytes, has been identified (36), and this protein is 54% similar to the Listeria YhdP protein (Fig. 5). Listeria contains one well-characterized sulfhydryl-dependent hemolysin, LLO, that is capable of lysing sheep erythrocytes (21). As L. monocytogenes hly mutants are nonhemolytic on blood agar plates, it is not likely that the yhdP locus encodes another Listeria hemolysin capable of red blood cell lysis under those conditions. Because of the homology of YhdP to hemolysin-like proteins, we constructed an in-frame chromosomal deletion of this gene in L. monocytogenes 10403s and tested its effect on virulence in mice. The yhdP mutant was attenuated approximately three- to sevenfold in mice compared to its wild-type parent (Fig. 6). This measured decrease in virulence, although small, was reproducible. Mice infected with the yhdP mutant looked noticeably healthier than those infected with the wild-type parent strain. It may be that YhdP shares overlapping roles with another protein as has been shown for the phospholipases PlcA and PlcB. Single mutations in either of these genes decrease virulence only slightly, while a plcA plcB double mutant is severely attenuated (41). Defining the function of the YhdP protein in Listeria awaits further experimentation.

Aside from actA, other previously identified L. monocytogenes in vivo-induced genes were not isolated. This could be due to the fact that the GFP plasmid library that was used is only a partial representation of the L. monocytogenes chromosome (see Materials and Methods). It is also possible that Sau3AI digestion created a bias against the formation of productive GFP fusions to other known in vivo-induced promoters. Libraries constructed with different restriction enzymes or DNases might identify additional genes when this method is used. Alternatively, it is possible that the isolation of actA promoters, but not other known in vivo-induced promoters, was so successful because of its high level of in vivo expression, a criterion that may be very important for the sorting step of this type of enhancement.

In conclusion, our results show that the FACS technique is a powerful tool for identifying potential in vivo-induced genes in gram-positive as well as gram-negative bacteria. The sequencing of the Listeria genome has been completed. This information should rapidly advance our understanding of Listeria pathogenesis. However, it is evident from these and other studies (14, 19, 24) that it is not always obvious which proteins are involved in the intracellular survival of a pathogen, and complementary biological studies are required to elucidate the complex interplay between a pathogenic microbe and its host.