![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Previous Article | Next Article 
Microbiology and Molecular Biology Reviews, June 2005, p. 217-261, Vol. 69, No. 2
1092-2172/05/$08.00+0 doi:10.1128/MMBR.69.2.217-261.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Centre of Microbial and Plant Genetics, Katholieke Universiteit Leuven, Kasteelpark Arenberg 20, B-3001 Heverlee, Belgium,1 Hogeschool voor Wetenschap & Kunst—De Nayer Instituut, Jan De Nayerlaan 5, B-2860 Sint-Katelijne-Waver, Belgium,2 School of Biological Sciences, University of Auckland, Private Bag 92019, Auckland, New Zealand,3 Department of Plant Sciences, University of Oxford, South Parks Road, Oxford OX1 3RB, United Kingdom4
SUMMARY INTRODUCTION IN VIVO EXPRESSION TECHNOLOGY Development of In Vivo Expression Technology Selection Strategies in IVET Auxotrophy-based selection. Antibiotic resistance-based selection. Recombinase-based selection. System-specific selection. Benefits and Shortcomings of IVET Strategies DIFFERENTIAL FLUORESCENCE INDUCTION Development and Applications Benefits and Shortcomings of DFI OVERVIEW OF IVET- AND DFI-ISOLATED GENES Genes Involved in Chemotaxis and Motility Genes Involved in Nutrient Scavenging Homeostasis of iron and other metal ions. Amino acid uptake. Acquisition of phosphorus. Uptake of sugars and carbohydrates. Miscellaneous nutrients. Genes Involved in Central Intracellular Metabolism Intermediary metabolic pathways. Lipid and fatty acid metabolism. Carbohydrate metabolism. Amino acid synthesis. Amino acid catabolism. Nucleotide synthesis. Protein synthesis and degradation. Cofactor biosynthesis. Genes Involved in Adaptation to Environmental Stresses Oxidative stress. Acid stress. Osmotic stress. Detoxification by efflux systems. Regulatory Genes Genes Involved in Cell Envelope Structure and Modification Peptidoglycan layer. Surface-exposed components. Outer membrane proteins. Genes Involved in Virulence and Secretion Type III secretion system. Virulence factors. Genes Involved in Nucleic Acid Metabolism Genes Involved in Transposition and Site-Specific Recombination FUN Genes CONCLUSIONS AND PERSPECTIVES IVET: a Powerful and Flexible Tool Concluding Remarks ACKNOWLEDGMENTS REFERENCES
| SUMMARY |
|---|
 Top NextReferences |
|---|
| INTRODUCTION |
|---|
|
TopPreviousNextReferences |
|---|
The utility of studying bacteria in vitro remains clear, particularly in light of technologies for genome-scale analysis in conjunction with the ability to carefully control biotic and abiotic environmental factors in the laboratory (184, 199, 238). For example, virulence factors of animal pathogens have been identified by analyzing bacterial responses to changes in temperature (166, 178), iron concentration (146, 182), pH (189), exposure to oxidative stress (282), and phosphate starvation (211). Similarly, the biology of rhizosphere-colonizing bacteria has been studied using simplified in vitro approaches. For example, root exudates have been collected from plant roots and used to study the bacterial response to root-derived factors (68, 176); the response of bacteria to other inhabitants of the rhizosphere has also been studied (294). Likewise, in vitro studies have proved useful for identifying host signal molecules triggering the onset of Agrobacterium tumefaciens pathogenesis (59, 61) and Rhizobium symbiosis (25, 207) upon interaction with plants.
Despite the value of in vitro studies, there is no escape from the fact that the vast majority of microbes exist in complex, dynamic environments that cannot be reproduced in the laboratory. For microbes, irrespective of their life style, there is growing recognition of the need to understand their function in the very environments that they inhabit and thus, ultimately, the causes of their ecological success.
Analysis of ecological success is far from straightforward: it is a complex and multidimensional phenotype determined by interconnected regulatory pathways involving both individual genes and gene networks. Natural selection, which is largely responsible for shaping the determinants of ecological success, does so by operating on interacting systems (more so than on single genes) to generate specific morphologies, physiologies, and behaviors. With this in mind, the value of different experimental approaches can be assessed.
Both bottom-up (genes to population) and top-down (population to genes) approaches have been used. The bottom-up approach is commonly used for studies of bacteria, although it is rarely pursued to the population level. The typical genes-to-phenotype strategy involves identification of traits on the basis of gene inactivation (143). This is a powerful approach that has been fundamental to the majority of advances in molecular microbiology, but, despite its power, insertional mutagenesis is not always appropriate for the analysis of phenotypes as complex as ecological performance. For most organisms, in most environments, there is no primary determinant of ecological performance; this is because it is determined by complex epistatic interactions among many different gene products that each have a long evolutionary history. Traits having the greatest effect on ecological performance are likely to be those that show subtle quantitative variation, and such traits are unlikely to produce "defective" phenotypes when inactivated (143).
Recent advances in gene fusion technologies provide an alternative way to study complex phenotypes. Rather than identifying genes on the basis of function loss, ecologically significant genes can be identified on the basis of their positive contribution to a specific phenotype. A study that aims to understand the mechanistic basis of ecological performance in bacteria colonizing a specific host might, therefore, begin by identifying those genes that are induced in the host environment. One advantage of this approach is that it considers bacteria as integrated organisms rather than as a toolbox of independent genes and phenotypes.
Bacterial gene expression can be determined by direct or indirect measurements of mRNA levels. Reporter gene fusions provide simple indirect methods for assaying transcription by placing a gene that encodes a product that can be readily assayed under the control of the promoter of interest. Two such reporters are lacZ (which encodes ß-galactosidase) (116) and gusA (which encodes ß-glucuronidase) (247). While reporters such as lacZ and gusA have been used most extensively to study gene expression in vitro, both of these reporters have also been used to study expression in complex environments, such as within the environment of living cells. However, improved reporters that encode luminescent (e.g., lux) or fluorescent (e.g., gfp) proteins have greatly increased the utility of transcriptional reporters to the extent that expression of single cells in complex environments can be studied (34, 41, 53, 85, 242).
In the past decade, many different techniques have been developed to study bacterial genes that are expressed during growth in specific and complex ecological niches (47, 138, 220). In this article we discuss the promoter-trapping techniques differential fluorescence induction (DFI) (279) and in vivo expression technology (IVET) (156), which have been used to identify and study genes showing elevated levels of expression in complex environments. In addition to the information these genes provide about the way that an organism perceives its environment, genes activated in a specific niche are likely to encode (or contribute toward) traits that are important determinants of ecological performance in that environment (201, 212). Complementary strategies such as signature-tagged mutagenesis (STM) (99), differential display using arbitrarily primed PCR (69, 172), subtractive and differential hybridization (111, 112), and selective capture of transcribed sequences (SCOTS) (86), are reviewed elsewhere (37, 94, 96, 154).
| IN VIVO EXPRESSION TECHNOLOGY |
|---|
|
TopPreviousNextReferences |
|---|
IVET (Fig. 1) is a promoter-trapping technique that selects microbial promoters active in a specified niche, for instance, during the interaction of a microorganism with its host. The first component of IVET is a conditionally compromised strain of the microorganism of interest that is mutated in a gene encoding an essential growth factor (egf) (220). The mutant strain is not able to sustain growth in the environment under study unless the egf gene is expressed. The second component of IVET is a plasmid carrying the promoter trap composed of the promoterless egf gene and a linked reporter gene (rep). Bacterial DNA is cloned randomly into the promoter trap and integrated in the chromosome of the egf mutant strain. Promoters that are specifically induced in the wild are identified by the ability to drive expression of the promoterless egf gene in this environment. This results in complementation of the mutation and, hence, in growth under the conditions encountered in the specified niche. To eliminate fusions with a "constitutive" promoter, recovered bacteria are screened for expression of the linked reporter gene (rep) on a general growth medium.
|
Bacteria harboring promoters that are specifically active in the wild are isolated from the specified niche, and the transcriptional fusions are rescued from the genome by standard molecular cloning procedures. However, this is laborious and can also be problematic. Therefore, alternative methods to recover fusions from the genome have been devised. One method is to recover the fusion by transduction using a suitable phage, e.g., bacteriophage P22 in the case of Salmonella spp. (155), but transducing phages are not widely available. A more generally applicable procedure to rescue chromosomally integrated plasmids is conjugative cloning (219). A helper plasmid supplies the genetic loci necessary for mobilization of the integrated plasmid into a suitable Escherichia coli host (220).
The IVET screening by Mahan et al. (156) relied on a purA or thyA null mutation, resulting in purine and pyrimidine auxotrophy, respectively, that greatly attenuated the growth of Salmonella enterica serovar Typhimurium during mouse infection (98, 156). Since then, a wide variety of genes encoding essential growth factors (Table 1), as well as different reporter genes have been used.
|
Auxotrophy-based selection. The auxotrophy-based selection strategy has been widely used in IVET screenings. All IVET studies based on this type of selection require a mutant strain defective in growth in the wild. This growth defect can be complemented by expression of the promoterless essential gene provided on the promoter trap.
As mentioned above, the first studies under the IVET moniker used auxotrophic Salmonella enterica serovar Typhimurium mutants defective in the de novo biosynthesis of purine or pyrimidine nucleotides combined with promoter traps supplying a promoterless purA or thyA gene, respectively, to complement growth of the Salmonella enterica serovar Typhimurium mutants in the wild (98, 156). The purA-based selection strategy was also used to identify Pseudomonas aeruginosa genes specifically induced during infection of mice (93). In another IVET study, purine auxotrophy of P. aeruginosa was obtained using a purEK mutant strain (289, 290).
Since not all bacteria show defective growth upon purA mutation and it is difficult to obtain a purA mutant strain for some microorganisms, several research groups used other essential genes. In principle, any biosynthetic gene that is necessary for growth in the wild can be used. The only prerequisite is that the auxotrophy cannot be complemented by metabolites retrieved from the occupied niche. Nevertheless, the gene to be mutated has to be chosen after careful consideration. For instance, when bacteria are able to reside intracellularly in host tissue, it has to be taken into account that nonsecreted host metabolites might also complement the auxotrophy.
Several authors adapted the auxotrophy-based selection strategy and used other essential metabolic genes (Table 1): panB, involved in pantothenate biosynthesis (218); dapB or asd, involved in diaminopimelate biosynthesis (63, 78, 95, 224, 245); metXW (159) or trpEG (26), necessary for methionine and tryptophan biosynthesis, respectively; inhA, required for mycolic acid biosynthesis (56); pyrB or thyA, necessary for de novo biosynthesis of pyrimidine nucleotides (140, 156); galU, involved in galactose metabolism (133); or ribBAH, involved in riboflavin biosynthesis (76). IVET was also applied to study infection of mice by the pathogenic fungus Histoplasma capsulatum. In this case, uracil auxotrophy was created by mutating the ura5 gene (226).
Antibiotic resistance-based selection. As it is not always easy or possible to construct an auxotrophic mutant, IVET selection based on antibiotic resistance, using antibiotic resistance genes as reporter genes (Fig. 2), is an important variant that expands the applications to a wider variety of microorganisms. While extending the utility of IVET, the use of antibiotic selection typically requires dosing the environment of interest with antibiotic, which inevitably changes aspects of the biological niche studied with implications for the spectrum of genes recovered.
|
Later, the cat gene was also used in IVET studies of Shigella flexneri (14), Salmonella enterica serovar Typhimurium (98, 157), Helicobacter pylori (8), Yersinia enterocolitica (300), Yersinia ruckeri (65), Streptococcus gordonii (127), Escherichia coli (126), and Burkholderia pseudomallei (239). Bacteria harboring promoters that are specifically induced in the wild were selected by administrating chloramphenicol to the host.
The use of antibiotic resistance genes as reporter genes in IVET studies is not limited to cat. To study Porphyromonas gingivalis virulence in mice, the promoterless tet gene was used, conferring tetracycline resistance (141, 298). Induced gene expression during pig infection by Streptococcus suis (254, 255) and mouse infection by Lactobacillus reuteri (285) was analyzed by an erythromycin resistance-based screening. And gene expression by Pasteurella multocida infecting mice was explored using a kanamycin resistance reporter gene (108).
Recombinase-based selection.
Both auxotrophy-based and antibiotic resistance-based selection have to cope with the inability to isolate transiently or weakly expressed genes. These disadvantages are circumvented by a second major modification of IVET screening, recombinase-based in vivo expression technology (RIVET). RIVET is based on the activation of a site-specific DNA resolvase and was initially used to identify Vibrio cholerae genes induced during infection of mice (32, 137). The resolvase used, TnpR from Tn
, is able to mediate recombination between two specific target sequences, the so-called res1 sites, and consequently slice out the interjacent DNA fragment from the genome.
In the first RIVET application, a tetracycline resistance gene (tet) was chosen as the reporter gene and was integrated into the chromosome, flanked by two res1 sites. The promoterless tnpR gene was provided on the promoter trap. Active promoters direct transcription of tnpR, and the activity of the resolvase results in excision of the reporter gene from the genome. By selection for tetracycline resistance during construction of the library, promoters that are active in vitro are discarded. After reisolation from the host, bacteria are screened for tetracycline sensitivity, and promoters active during interaction with the host are retained. The RIVET strategy has also been validated to study Staphylococcus aureus infection of mice (152). In this case, a kanamycin resistance gene was used as reporter gene and was integrated into the chromosome, flanked by two res1 sites.
A similar system was used in a RIVET strategy to study Salmonella enterica serovar Typhimurium infection of mice (5). This system consists of a promoterless derivative of cre, encoding the phage P1 recombinase, carried on the promoter trap. The targets of the Cre recombinase are two chromosomally integrated loxP sites flanking the npt gene, conferring kanamycin resistance.
RIVET is applicable to many microorganisms, even those that are difficult to manipulate since only the reporter gene flanked by recognition sites has to be integrated into the chromosome. A Cre-based RIVET system was devised by Bron et al. (24) for lactic acid bacteria. To study infection of mice by the fungal pathogen Candida albicans, Staib et al. devised a RIVET system consisting of an Flp recombinase and a genetic marker, conferring resistance to mycophenolic acid, flanked by the specific recognition sites for the recombinase (262).
System-specific selection. It can be of interest to identify genes (promoters) differentially expressed during a particular stage of the interaction between a bacterium and its eukaryotic host. To this end, dedicated IVET strategies can be devised in which the promoter trapping gene encodes an "essential interaction factor" (eif) required at a specific stage of the interaction. Bacteria are screened for the ability to establish a firm interaction with the host. It is therefore necessary that the establishment of the microbe-host interaction result in a scorable host phenotype, such as cell lysis, plant disease, or symbiosis.
For example, a specific IVET selection strategy was devised to isolate Sinorhizobium meliloti genes that are specifically induced in the early stages of symbiosis (188). In this study, the bacA and gusA genes were used as reporter genes to assess host-induced and in vitro promoter activity, respectively. BacA is an integral membrane protein that affects the degree of modification of the lipopolysaccharides. BacA is required for intracellular infections during Sinorhizobium meliloti plant symbiosis and Brucella abortus animal pathogenesis (64). BacA is also necessary for differentiation of Sinorhizobium meliloti into nitrogen-fixing differentiated cells (bacteroids) (109). Only when active promoters are inserted in the promoter trap will the bacA gene be expressed, resulting in the differentiation process. Nitrogen-fixing nodules containing bacteroids can readily be distinguished from non-nitrogen-fixing nodules by macroscopic observation. In this way, the screen targets genes that are expressed after the initiation of nodulation but before bacteroid differentiation and nitrogen fixation take place. Isolation of genes known to be involved in nodulation (e.g., nifS) suggests that the strategy functions as expected. Moreover, it enabled identification of genes that were not previously associated with nodulation (188).
A similar strategy was developed to isolate genes involved in the early stage of Pseudomonas syringae pv. tomato infection of Arabidopsis thaliana leaves (16). This IVET consists of a Pseudomonas syringae hrcC mutant strain with a deficient type III secretion system (TTSS). TTSS is necessary for infection and growth in susceptible plants. Subsequently, a promoterless hrcC gene was used as the reporter in the promoter trap. Only genes expressed during establishment of infection can be isolated with this modified IVET. The approach used here proved useful, since 40% of the transcriptional fusions revealed genes already known to be involved in pathogenesis. Validation was obtained by isolation of hrp/hrc and avr genes, encoding proteins of the TTSS, as they are known to be induced upon inoculation and hence during the early stage of infection. A similar system is being developed with a Xanthomonas campestris pv. vesicatoria hrpB1 mutant also defective in expression of a functional TTSS (U. Bonas, personal communication).
To isolate Listeria monocytogenes virulence genes, a modified IVET was devised based on hly, encoding a hemolysin (listeriolysin) (55, 77). Listeriolysin is a virulence factor absolutely required for intracellular survival and growth in mice. Disruption of hly results in the loss of the hemolytic phenotype on blood agar plates and a severe decrease in virulence. Consequently, infection of mice by hly mutants can only occur when the hly gene provided on the promoter trap (lacking its cognate promoter) is expressed. After isolation of the infecting bacteria, the same reporter gene (hly) was used to screen promoter activity in vitro, since hemolysis is apparent as a zone surrounding Hly+ bacteria on blood agar plates. Again, this modified IVET focuses on genes that are induced and necessary in the early stages of infection rather than genes that enable bacteria to adapt to and survive in the new environment.
In an IVET application to study the plant pathogen Pseudomonas syringae pv. syringae, a methionine auxotrophy-based selection strategy was devised. In moist plant leaves, the metXW mutant used displays normal growth, but shows severely attenuated growth on plants in dry conditions (159). In this way, the auxotrophy-based selection only occurs when plants are transferred to dry growth conditions, and the timing and degree of selection pressure can be altered accordingly. For instance, by growing the plants in wet conditions in the early stages of infection, the conditionally compromised metXW mutants are able to grow and establish large populations. The IVET selection regimen is subsequently started by transferring the plants to dry conditions. Therefore, the name habitat-inducible rescue of survival was introduced (159).
The green fluorescent protein (GFP)-based IVET leaf array for identification of plant-upregulated genes in Erwinia chrysanthemi, described by Yang et al. (299), does not involve positive in planta selection using an essential growth factor gene. Selection of induced promoters is based upon differences in fluorescence intensity during plant infection and during growth on a general growth medium.
Since the early IVET studies of animal infection by Salmonella enterica serovar Typhimurium and Pseudomonas aeruginosa, IVET has been adapted and applied to study a wide variety of microorganisms. It is clear from Table 1 that the various IVET selection strategies extended its use to study differential gene expression not only in gram-positive bacteria but also in eukaryotic microorganisms such as Candida albicans (262) and Histoplasma capsulatum (226).
IVET is not technically demanding and can be applied using standard molecular biology techniques. This means that in contrast to DFI or microarrays, no expensive equipment is required. Another major advantage of IVET is that no extensive knowledge of the genome of the microorganism under study is required to apply the technique. For example, Yersinia ruckeri and Pseudomonas stutzeri A15 are bacteria for which only a few DNA sequences were analyzed in the past, but IVET proved a useful technique to analyze gene expression in their host environments (65, 224). However, the availability of a (draft) genome sequence of the target microorganisms or close relatives certainly speeds up subsequent characterization of the trapped promoters.
Variations on the original IVET theme (antibiotic resistance-based IVET, RIVET, and DFI) have enabled the study of microorganisms for which straightforward genetic analysis, such as construction of defined mutants, is not readily available. For instance, in eukaryotes the presence of two alleles for each gene hampers mutational analysis, but IVET techniques enabled gene expression analysis of pathogenic fungi in their natural habitat (226, 262).
IVET can be applied to microorganisms residing in ecological niches that are very different in nature. IVET has been successfully applied to study microorganisms residing in animal hosts (ranging from fish, pigs, and chinchillas to mice), in macrophages, in plants, in the rhizosphere, and even in bacteria colonizing an oomycete (Table 1). When the microorganism under study is able to colonize two different host organisms, host-triggered gene expression can be assessed using the same promoter trap library. Comparing the two (different) subsets of host-induced genes provides information about the differences and similarities in the microenvironment of the two host organisms.
IVET is not limited to studying interactions with animal, fungal, or plant hosts, but can be extended for use in other complex environments. IVET was, for instance, used to study P. aeruginosa in biofilms with the so-called in-biofilm expression technology (67), or to study differential gene expression of P. aeruginosa during infection of burned mice tissues (92). A dapB-based IVET system was used to explore the genetic needs for survival of Pseudomonas fluorescens Pf0-1 in bulk soil (245). An IVET technique is also being developed to study gene expression of the oil-degrading marine bacterium Alcanivorax borkumensis in response to key environmental signals in order to study the bacterial determinants involved in biodegradation of hydrocarbons (82). In addition, an IVET-like strategy has been used to study differential gene expression in different genetic backgrounds. With the so-called identification of transcriptional regulator-activated promoters, the dependence of the transcription of Mycobacterium tuberculosis genes on various transcriptional regulators such as sigma factor 
E could be analyzed by monitoring the reporter gene activity in a E-overexpressing and a E knockout strain of Mycobacterium smegmatis (173).
IVET has many attractive features, but some possible drawbacks have to be considered in the interpretation of the resulting data. First, IVET is not designed to isolate repressed promoters. Second, the subset of genes that are identified depends on the strength of the selection regimen in the wild. In each experimental system, the strength and the method of selection in the wild have to be chosen with consideration. If the selection is too strong, weakly or transiently expressed promoters will not be identified and highly expressed genes will be favored in the screening. On the other hand, a too weak selection in the wild will lead to false positive results. Third, proteins that are expressed constitutively but only activated in the wild (for example, by phosphorylation) are not detected. Fourth, the sets of genes defined as specifically induced in vivo are partially dependent on the "in vitro" growth conditions used to assess whether the reporter gene fusion is inactive outside the environment under investigation. For instance, the composition of the growth medium can significantly impact the expression of genes involved in nutrient acquisition and metabolism. Finally, mutants affected in genes that are isolated with IVET have to be constructed and phenotypically characterized. Only for a limited number of model microorganisms are ordered mutant libraries available that cover the entire sequenced genome.
As IVET can, in principle, be applied to study virtually all culturable microorganisms in their complex environments, it is clear that many researchers benefit from using the IVET strategy to study their favorite bug. The choice of selection strategy is facilitated by the development of the different IVET modifications. However, each specific selection strategy comes with its own advantages and disadvantages.
The major disadvantage of autotrophy-based selection is the need to construct an auxotrophic mutant, and for some microorganisms the tools to achieve this are not (yet) available. However, the nature of the auxotrophic mutation can determine in part the strength of the selection in the wild. When the generated auxotrophy is lethal for actively growing cells but does not impair cell survival, auxotrophy-based selection becomes a very powerful tool since the strength of selection in the wild can be easily adjusted by altering the time lapse between infection and reisolation (220). Whether low or transiently expressed genes will be detected depends on the strength of the selection regimen.
Switching to antibiotic selection avoids the construction of a mutant strain, thereby increasing the applicability of IVET. However, drug administration to the host might interfere with the complex process of interaction, e.g., with the immune defense of the host. Due to the presence of antibiotics, the composition of the natural ecological niche of which the microorganism is part might be altered. Furthermore, antibiotic administration to the host is not always possible, as the host organism itself might be affected by antibiotic treatments, as is often the case with plants (190). Moreover, to study microorganisms that reside within plant tissues, this selection strategy is scarcely suitable since several antibiotics are not translocated to all plant tissues. Once a suitable antibiotic for selection is found, it is important to evaluate the proper dose of antibiotic administration to allow selection of promoters that drive expression of the antibiotic resistance gene. The selection regimen in the wild can be modified easily. Variation of the antibiotic concentration allows isolation of genes that are expressed at different levels. Changing the time of drug administration permits isolation of genes that are expressed at different time points.
The main disadvantage of RIVET is the loss of a positive selection strategy after reisolation from the environment. This screening is rather laborious since isolated microorganisms are tested for antibiotic sensitivity by replica plating. However, Merrell and Camilli (175) solved this problem by inserting, together with an antibiotic resistance gene, a second reporter gene, sacB, into the excisable cassette. Its gene product, levansucrase, catalyzes conversion of sucrose into levan, which is toxic for most gram-negative and some gram-positive bacteria and results in defective growth in media containing sucrose. The sacB reporter gene can be used for a positive selection because bacteria that contain promoters induced in the wild have lost the reporter gene cassette, thereby enabling growth on media containing sucrose. Another strategy to avoid negative selection with RIVET is the use of a cat resistance gene which is disrupted by a tet resistance gene flanked by res sites (147). With this so-called selectable in vivo expression technology, the nonresolved strains remain resistant to tetracycline, while the resolved strains become resistant to chloramphenicol.
The major advantage of RIVET compared to antibiotic- and auxotrophy-based selection consists in the isolation of weakly or transiently expressed genes. However, the sensitivity of RIVET could also turn into a disadvantage, as genes important in the wild but expressed in vitro at a moderate to high basal level will not be isolated. The gene of study must be transcriptionally silent during strain construction and propagation in vitro. Otherwise, the antibiotic cassette is spliced out during the library construction, and consequently, the bacteria cannot survive the antibiotic selection. For this reason, fine tuning of the RIVET selection strategy was achieved by modulating the ribosome binding site of the promoterless recombinase gene (139). As a result of mutations in this ribosome binding site, translation efficiency at any transcriptional level is decreased, resulting in less sensitive selection.
Promoter traps consisting of tnpR alleles with different translation efficiencies render different pools of isolated genes that are induced to some level in the wild due to the lowered sensitivity of RIVET (139). Recently, a tunable RIVET system to study Vibrio cholerae infection was also achieved using resolvable cassettes with different efficiencies of excision (191). However, for each tnpR allele or resolvable cassette, a separate fusion library has to be constructed, which multiplies the manipulations.
RIVET permits the analysis of spatial and temporal gene expression (139). The expression patterns of the genes of interest can be investigated at different time points of the interaction, at different anatomic sites of the host organism, and even in different hosts or different genetic backgrounds simply by determining the proportion of resistant bacteria versus those sensitive to the antibiotic.
| DIFFERENTIAL FLUORESCENCE INDUCTION |
|---|
|
TopPreviousNextReferences |
|---|
DFI was originally designed to isolate acid-inducible genes in Salmonella enterica serovar Typhimurium (278, 280). This technique was subsequently adapted to study induction of Staphylococcus aureus and Streptococcus pneumoniae gene expression by in vitro stimuli that mimic the host environment, such as temperature shift, increased osmolarity, iron limitation, increased acidity, presence of competence stimulatory peptide, and cation starvation (13, 160, 237).
DFI studies of Salmonella enterica serovar Typhimurium, Mycobacterium marinum, and Listeria monocytogenes showed that DFI is not limited to studies of the effect of in vitro stimuli mimicking the host, but also enables analysis of gene expression during infection of macrophages (11, 221, 279, 295). Although these studies demonstrated successful sorting of macrophages based on the fluorescence of infecting bacteria, it is worth noting that the bacterial population within an infected macrophage is heterogeneous, which might lead to false positive results. However, it was possible to apply DFI to study Streptococcus pneumoniae infection when the pathogen was isolated from host body fluids, such as blood (160). Using two-color flow cytometry, Bumann (30) successfully analyzed gene expression of Salmonella organisms isolated from mouse Peyer's patches. Recently, the use of DFI was extended to explore differential gene expression of plant-associated bacteria such as Rhizobium leguminosarum (3), Pseudomonas syringae (35), and Bacillus cereus (57).
In contrast to IVET, transcriptional fusions are not integrated into the chromosome but are provided on plasmids because single-copy gfp expression results in sufficiently intense fluorescence for accurate measurement only when driven by a strong promoter. In addition, the use of plasmids facilitates the isolation of in vivo-induced promoters. However, the use of multicopy plasmids prevents detection of context-dependent or topology-dependent effects of gene regulation.
DFI shares with IVET the caveats that are inherent to promoter trap approaches, such as the inability to detect genes that are regulated posttranscriptionally and the need to construct and analyze mutated target genes to assess their role in the wild (154). However, DFI shows additional disadvantages intrinsic to the technology. Flow cytometric analysis and sorting can be hampered by aggregation of bacteria or macrophages. Additional problems arise with the isolation and fluorescence quantification of bacteria that are isolated from infected host tissues because of the prevalence of background fluorescent particles (138, 278).
Other disadvantages are associated with the use of GFP, such as restrictions to the pH range in the studied environment, oxygen requirement for fluorophore development, and the absence of signal amplification. Due to the nonlinearity of fluorescent signals, it is necessary to calibrate and determine the linear range of the signal for each experiment to allow quantification of gene expression (180). However, most of these problems have been solved by the technological advances made concerning maturation, fluorophore development, fluorescence intensity, and spectral properties of GFP (42, 43, 258, 276).
| OVERVIEW OF IVET- AND DFI-ISOLATED GENES |
|---|
|
TopPreviousNextReferences |
|---|
In Table 2, host-induced genes identified with IVET or DFI are classified in 9 functional groups. Most of the data in Table 2 were obtained from IVET studies of bacterial pathogens during infection of a mammalian host. In most cases a murine infection model was used to explore host-induced gene expression, but the interaction of animal pathogens with fish, pigs, rabbits, and chinchillas was also studied with promoter traps. In recent years, however, the IVET technique has found wider application for exploration of in planta gene expression of phytopathogenic bacteria as well as in nonpathogenic systems of bacterial interaction with crop plants (alfalfa, sugar beet, rice, and maize). In the majority of the studies, -proteobacterial members (17 species) were covered, mainly enterobacteria (seven species) and Pseudomonas (five species). Among the gram-positive bacteria, most data originate from members of the firmicutes (six species), whereas the use of IVET to study actinomycetes has only been reported for Mycobacterium tuberculosis. The data of 11 DFI studies of microbe-host interactions are incorporated in Table 2. DFI was used almost exclusively to study animal pathogens, mostly belonging to the -Proteobacteria (three species), firmicutes (three species), and mycobacteria (two species).
|
25) of the FliF protein are assembled into the cytoplasmic membrane-embedded MS-ring, which constitutes the core of the flagellar motor (268). The isolation of these fliF genes indicates that the flagellar machinery is important for gram-positive as well as for gram-negative bacteria to establish interactions with both animal and plant hosts. This conclusion from IVET studies is in line with the results from other studies. For efficient colonization of tomato roots by nonpathogenic Pseudomonas fluorescens, flagellum-driven chemotaxis is required (48). In several other cases where competition between several bacterial species exists, flagellum-mediated motility is shown to provide a specific advantage for bacteria (181). For instance, for Vibrio cholerae, it was shown that mutants with knockouts in a flagellar subunit gene (flaA) or in genes encoding flagellar motor proteins (motAB and motY) are severely affected in colonization of the mouse small intestine (137).
Besides providing motility, flagella are important for bacterial attachment to surfaces and are thus generally considered important virulence factors (45, 181). Therefore, it is not surprising that the flagellar subunit is recognized by the innate immune system in organisms as diverse as plants (306) and mammals (256). This implies that for a successful host infection to occur, a bacterial pathogen might need to suppress flagellar assembly after the initial interaction stage. This requires strictly coordinated temporal gene expression during the different stages of interaction. Antisense transcripts of target genes may be involved in this. A putative antisense 
-flaA was identified in Vibrio cholerae by RIVET (32). Likewise, a putative antisense transcript of fliM, encoding the flagellar C-ring, is upregulated in Pseudomonas stutzeri A15 during rice root colonization (H. Rediers, unpublished data). It is known that some bacteria adapt their flagellation pattern in response to the environmental conditions they encounter (reviewed in references 73 and 170).
The apparent temporal expression/repression of flagellin synthesis may be coordinated with the expression/repression of the cognate chemosensory machinery (Che system). A cheY-containing transcriptional fusion was isolated by an IVET screening in Pseudomonas aeruginosa infecting mice (290) and Ralstonia solanacearum infecting tomato plants (26). CheY is the response regulator protein that, in its phosphorylated form, interacts with the switch machinery of the flagellar motor to change the direction or speed of rotation (269). In addition to the temporal regulation of flagellin genes, chemotaxis may be fine-tuned throughout the infection process by differential expression of methyl-accepting chemotaxis protein subsets and multiple motility-linked chemosensory systems that are present in many bacteria, such as Vibrio spp. (169), Pseudomonas spp. (66), and Rhodobacter spp. (163). In two IVET studies, putative antisense transcripts of chemotaxis-related genes, encoding both sensory (-mcp) and signal transduction (-cheV) proteins were identified in rice-colonizing Pseudomonas stutzeri (224) and Vibrio cholerae infecting mice, respectively (32). CheV encodes a chimeric protein containing a CheY-homologous domain. It has been shown that a Vibrio cholerae cheV mutant colonizes mice better than the wild type (32). This is in agreement with the reciprocal regulation of motility and virulence genes in Vibrio cholerae (80). Downregulating chemotaxis genes might increase infection efficiency by favoring the formation and maintenance of microcolonies (175).
IVET also revealed that a gene involved in type IV pilus biogenesis is upregulated in Ralstonia solanacearum during tomato infection (26). Type IV pili enable twitching motility, a pilus-based form of translocation used by pathogens to spread over the host tissue surface, and are therefore recognized as important virulence factors for a wide range of plant and animal pathogens. In addition, type IV pili are important for the formation of biofilms and fruiting bodies (165).
-proteobacterial species in a plant or animal host environment. After iron chelation, ferrisiderophores are captured at the cell surface by specific high-affinity siderophore receptors. Consistent with the induction of siderophore biosynthesis, several genes encoding such receptors have been identified by IVET or DFI. The ferric hydroxamate receptor, encoded by fhuA, is induced in both Salmonella enterica serovar Typhimurium infecting mice (98) and Shigella flexneri infecting monolayers of human epithelial cells (233). In Porphyromonas gingivalis, a putative siderophore receptor encoded by ivi10 is specifically induced during infection of mice. Wu et al. (298) demonstrated that a Porphyromonas gingivalis ivi10 knockout mutant is outcompeted by the wild type during survival in the host and displays a reduced ability to cause infection. The active translocation of ferrisiderophores is Ton dependent and is driven by the proton motive force. Following TonB-dependent translocation, ferrisiderophores are finally transported into the cytoplasm by an ATP-binding cassette (ABC) transporter. Genes encoding the RupDGC transporter as well as other genes involved in ruckerbactin-mediated iron uptake (synthesis of siderophore and receptor and TonB-dependent translocation) were identified by IVET in Yersinia ruckeri infecting fish (65).
Upon infection, several pathogenic bacteria display upregulation of genes encoding siderophore-independent iron uptake systems, such as the Salmonella enterica serovar Typhimurium SitABCD transporter (305). Application of IVET revealed that the sitABCD operon is specifically induced during infection of mice. It was subsequently demonstrated that a Salmonella enterica serovar Typhimurium sit mutant is severely attenuated in infection of mice (114). The sitABCD operon encodes an ABC transport system that mediates iron and probably also manganese uptake (124). Erwinia chrysanthemi yfeA, which encodes a component of a Sit-homologous transport system, is upregulated during plant infection (299). Other promoter trap studies demonstrated that sitA and sitC homologues are specifically expressed in Shigella flexneri (233) and Yersinia enterocolitica (83), respectively, during infection of mice. Although a sitA mutation does not affect plaque formation by Shigella flexneri on monolayers of human intestinal epithelial cells, a sitA mutation in combination with other iron acquisition mutations shows additive effects in these plaque assays (234).
Some bacteria possess mechanisms for the uptake of siderophores that are produced by other species or for uptake of iron-containing host proteins such as transferrin or heme (232, 251). IVET revealed the upregulation of genes involved in hemin uptake (hmuS, hmuT, and hmuU) in spinach-infecting Erwinia chrysanthemi (299). The Burkholderia pseudomallei hemT gene, encoding a periplasmic hemin binding protein, was also shown to be induced during macrophage infection (M. S. Thomas, personal communication).
In Pseudomonas aeruginosa, the np20 gene, encoding a homologue of the ferric uptake regulator (Fur), was also found to be specifically expressed during infection of mice. Mutational analysis revealed that np20 is not essential for growth in vitro. However, compared to the wild type, the mutant strain is required in a much higher dose to cause similar lethality in mice (290). Besides regulation of iron uptake, the housekeeping Fur protein is also directly or indirectly involved in the regulation of a substantial number of other genes encoding proteins with remarkably diverse functions, including other regulators, proteins involved in adaptation to oxidative stress, and virulence factors such as exotoxin A (284). Likewise, it was shown that in the plant pathogen Erwinia chrysanthemi, besides expression of two high-affinity iron uptake systems, pectate lyase-mediated cell wall degradation is also under control of the Fur regulator (72).
The frequent isolation of genes related to iron homeostasis reflects the importance of iron for microbial growth. Animal pathogens reside in an environment low in iron ions because host proteins such as transferrin and lactoferrin bind iron with high affinity (236). These proteins also play a role in host protection against microbial infection at the mucosal surface by depletion of the available iron (291). Likewise, in certain soils, the plant rhizosphere is scarce in ferrous iron (149, 195). Because iron is essential for microbial growth, the animal pathogens and plant-associated bacteria deploy dedicated systems for high-affinity iron uptake to capture the available iron (287). Furthermore, in several plant and animal pathogens, expression of pathogenesis-related genes is linked to iron availability (72, 206, 228, 283, 284).
Several genes involved in scavenging other metal ions, such as Cu2+, Mn2+, Mg2+, K+ and Na+ were identified with IVET. Using DFI, the Streptococcus pneumoniae psaBCA operon, encoding a manganese uptake system, was identified as being specifically expressed in mice (160). The psa promoter is induced more than 10-fold, suggesting an important role during survival within the host. Moreover, psaB, psaC, and psaA mutants are not only growth retarded in medium low in manganese, but are also completely attenuated in infection of mice (161). The Streptococcus pneumoniae Psa permease also plays an important role in resistance to hydrogen peroxide and superoxide, in systemic infections, and in nasopharyngeal mouse colonization (168).
In Sinorhizobium meliloti, a potassium channel, encoded by nex10, is specifically induced in the nodules during alfalfa symbiosis. A nex10 mutant strain is less efficient in symbiotic nitrogen fixation (188). The nex10 gene product resembles the regulatory ß-subunit of the eukaryotic voltage-gated potassium channels, but the exact function of this type of channel in prokaryotes is unknown. Possible roles in osmoregulation or pH adaptation during symbiosis have been suggested (177, 188).
Amino acid uptake. Although most IVET and DFI studies focused on animal infection systems, genes involved in amino acid acquisition were predominantly isolated from plant-associated bacteria, suggesting that amino acids are available for uptake in the plant environment but much less so in animal hosts. The host-induced expression of genes for amino acid uptake systems was reported for pathogenic bacteria upon plant infection (16, 26) as well as for beneficial bacteria colonizing the plant rhizosphere (218). Part of the amino acids synthesized by plants is exuded into the rhizosphere (113). Although the amount of amino acids in tomato root exudates is not sufficient to sustain rapid growth of plant root-colonizing microorganisms, the amino acids are likely to be taken up and utilized during colonization (249). This is supported by the observation that Pseudomonas fluorescens shows chemotaxis towards amino acids present in these root exudates (48).
Acquisition of phosphorus. IVET and DFI studies have revealed host-induced expression of systems for uptake of phosphorus in both animal- and plant-pathogenic bacteria. For instance, the Shigella flexneri pstS encodes an ABC transporter for high-affinity phosphate uptake that is specifically expressed during infection of mice. The Shigella flexneri pstS mutant strain constructed shows no growth difference in low-phosphate media but causes smaller plaques on macrophage monolayers, suggesting that the loss of pstS results in lower infection efficiency (233).
Uptake of sugars and carbohydrates. Several bacterial sugar uptake systems, mostly sugar permeases and sugar-specific phosphotransferase systems (PTS) for monosaccharides (fructose, mannose, and ribose) and disaccharides (sucrose, cellobiose, and maltose), were found to be induced during interaction with several eukaryotic hosts. During passage through the mouse gastrointestinal tract, Lactobacillus plantarum seems to deploy a diverse set of PTS systems for uptake of sugars (24). Sugar uptake systems specifically associated with plant infection were revealed for Erwinia chrysanthemi (299). Mutation of rhiT, encoding a rhamnogalacturonide transporter, compromises the systemic invasion capability of this phytopathogen (299).
Three genes (dctS, dctD, and the dctS homologue Rsc1598) involved in the regulation of C4-dicarboxylate uptake were isolated with independent IVET screenings for the rhizosphere-colonizing Pseudomonas fluorescens (218) and from the phytopathogens Pseudomonas syringae (159) and Ralstonia solanacearum (26), respectively. C4-dicarboxylate metabolism is induced in the presence of dicarboxylates and is under the control of regulatory sensor mechanisms. The DctSR two-component regulatory system, which shows high similarity with the FixLJ oxygen sensor system in rhizobia, is necessary for aerobic growth on C4-dicarboxylates. The DctBD two-component regulatory system is functionally similar to the NtrBC regulatory system that activates expression from 54-dependent promoters (115). In a similar, ATP-dependent way, DctBD activates expression of dctA, encoding a C4-dicarboxylate:cation (H+ or Na+) symporter, which is essential for symbiotic nitrogen fixation (115). C4-dicarboxylates such as malate and succinate are present in plants and root exudates (9) and are major carbon and energy sources for nitrogen-fixing symbionts (301). IVET studies highlight the significance of this catabolic pathway for other plant-associated bacteria as well.
Miscellaneous nutrients. IVET studies have revealed a variety of other genes specifically expressed in the wild, encoding ABC transporters, porins, and permeases for uptake of diverse components, such as lactate, peptides, choline, and undefined molecules. The potential role of an ABC transporter (ATP-binding subunit RTI006) of Streptococcus pneumoniae was revealed by DFI (160). This transporter mediates choline transport and subsequent analysis showed that the Streptococcus pneumoniae RTI006 mutant strain shows decreased respiratory tract infection (160). In Staphylococcus aureus, choline and its degradation product, glycine betaine, constitute potent osmoprotectants (87, 231). Besides their involvement in uptake of nutrients, substrate-binding components of ABC transporters have also been shown to be implicated in the infection process of some pathogens by facilitating adhesion to host cells, as unequivocally demonstrated in Streptococcus gordonii (117) and Campylobacter jejuni (205).
IVET also demonstrated that some TCA cycle genes are upregulated in the host environment. The TCA cycle is a major degradation pathway for generation of ATP but also provides intermediates for biosyntheses. For instance, fumC, encoding the fumarase enzyme, showed elevated expression in Listeria monocytogenes during infection of mice (77). Analysis of a Listeria monocytogenes fumC mutant strain revealed defective growth in phagocytes (77). It is worth noting that fumC was also identified in the IVET screening of Ralstonia solanacearum upon infection of tomato plants (26).
Reduced coenzymes produced by oxidative metabolism, such as the TCA cycle and fatty acid degradation, can be used to drive ATP synthesis via oxidative phosphorylation. The ATP synthase subunit gene atpD was isolated with IVET in a Staphylococcus aureus mouse infection model (152). Another gene involved in energy metabolism, pckA, was found to be up-regulated during Sinorhizobium meliloti-alfalfa symbiosis (188). The phosphoenolpyruvate carboxykinase PckA catalyzes the formation of phosphoenolpyruvate from oxaloacetate. It was previously shown that rhizobial pckA is strongly induced at the onset of stationary phase or during growth on succinate or arabinose as the sole carbon source, but pckA expression is also induced by host root exudates (193). A Rhizobium pckA mutant strain revealed a host-dependent symbiotic phenotype, as it lost its ability to establish nitrogen-fixing nodules only in some leguminous plants (194). Using IVET, pckA was also found to be induced in Mycobacterium tuberculosis during macrophage infection. The host-induced expression of pckA was subsequently confirmed with reverse transcription-PCR (56).
Lipid and fatty acid metabolism. IVET, RIVET, and DFI screenings revealed that several microbial genes involved in lipid and fatty acid metabolism are upregulated in animal and plant host environments. Degradation of lipids from host immune cells might protect the pathogen against the host immune response, but, as outlined above, lipid degradation and subsequent utilization of the released fatty acids may also fulfill a nutritional role (32, 157). It was also proposed that pathogen lipases, in combination with fatty acid-modifying enzymes, could inactivate the bactericidal lipids that are produced in host tissue abscesses, thereby increasing survival in this niche (121). This is in agreement with the isolation of the Staphylococcus aureus lip gene, encoding a glycerol ester hydrolase, which is specifically induced in host tissue (152).
The upregulation of genes encoding a 3-hydroxyl-coenzyme A dehydrogenase in Mycobacterium tuberculosis infecting macrophages was demonstrated independently with IVET (56) and DFI (275). An unexpectedly large number of host-induced genes that are involved in fatty acid metabolism were identified by IVET in Mycobacterium tuberculosis, suggesting that fatty acid metabolism is extremely important during life in the host environment. Mycobacterium tuberculosis has an astonishingly large number of genes presumed to be involved in ß-oxidation of fatty acids (39). It has been suggested that Mycobacterium tuberculosis is able to utilize fatty acids as a major energy source during infection, but it is also possible that fatty acid metabolism is necessary for remodeling the cell envelope upon macrophage infection, thereby evading the host immune response (56). Likewise, Mahan and coworkers found that fadB, which is required for ß-oxidation of fatty acids, is upregulated in Salmonella enterica serovar Typhimurium upon infection of mice and ascribed this to the high concentration of fatty acids encountered by the pathogen during infection (157). FadB homologues were also found to be specifically expressed in Brucella abortus and Mycobacterium tuberculosis (62, 275).
In the plant pathogen Erwinia chrysanthemi, the transcriptional regulator EutR was isolated using IVET (299). This regulator controls the eut operon, which is required for ethanolamine utilization (243). Ethanolamine utilization is an important trait for efficient plant infection, since an Erwinia chrysanthemi eutR mutant displayed a decreased ability to cause systemic invasion in African violets (299). Ethanolamine, released during degradation of phospholipids, is also a carbon and energy source in the intestinal tract, and the eut operon is coregulated with expression of virulence genes (125, 135). In Salmonella enterica serovar Typhimurium, the operon involved in ethanolamine utilization is coregulated with genes involved in motility and with genes encoding a TTSS (125).
Carbohydrate metabolism. Several of the genes listed in Table 2 are involved in sugar metabolism. One of these genes, xylA, encoding xylose isomerase, was identified in the sugar beet colonizer Pseudomonas fluorescens (218). Xylose is a typical plant-derived sugar commonly present in the plant rhizosphere. Xylan is the main carbohydrate found in the hemicellulosic fraction of plant tissues and is hydrolyzed by xylanases into xylose monomers that can be utilized by many bacteria and fungi as a primary carbon source. Xylanase producers are found in all ecological niches where plant material is deposited (210). It is therefore plausible that genes enabling Pseudomonas fluorescens to utilize xylose are switched on when the organism is residing in the rhizosphere. It is known that the XylR transcriptional regulator activates xylA expression in the presence of xylose (257). The IVET isolation of xylR in the plant pathogen Erwinia chrysanthemi indicates that xylose metabolism is also activated in planta (299). Notably, xylA was also isolated in the mouse gut colonizer Lactobacillus reuteri (285). We speculate that Lactobacillus reuteri xylA might be induced during survival in the gut by xylose originating from plant material in the mouse feed.
It was mentioned above that promoter traps enabled the isolation of several sugar uptake systems, such as the ribose transport proteins RbsD and RbsC, from Lactobacillus plantarum during passage in the gastrointestinal tract of mice (24) and Haemophilus influenzae infecting chinchilla, respectively (164). Likewise, other members of the ribose (rbs) operon, rbsR and rbsK, were isolated with IVET from Lactobacillus plantarum (24) and from Klebsiella pneumoniae (133), respectively, during life in the mouse host. RbsR is a transcriptional repressor of the rbs operon, which enables uptake and utilization of ribose. The rbsR gene is also upregulated in Salmonella enterica serovar Typhi during macrophage infection, as demonstrated with SCOTS (44). An rbsR mutant exhibits decreased survival in macrophages compared to the wild type, confirming the importance of rbsR induction during macrophage infection (44). Differences in the regulation (repression versus activation) of ribose metabolism in the different host environments might be explained by different sugar contents present in the spleen and intestine or experienced inside macrophages.
Amino acid synthesis. The importance of amino acid uptake for bacterial life in the plant environment was discussed. Likewise, amino acid synthesis seems to be a key trait for interaction and/or survival in the host environment. Members of this class were found to be induced in microorganisms as diverse as the fungal pathogen Histoplasma capsulatum and gram-positive and various gram-negative
