ABSTRACT
Two broad-host-range vectors previously constructed for use in soil bacteria (A. G. Matthysse, S. Stretton, C. Dandie, N. C. McClure, and A. E. Goodman, FEMS Microbiol. Lett. 145:87–94, 1996) were assessed by epifluorescence microscopy for use in tagging three marine bacterial species. Expression of gfp could be visualized in Vibrio sp. strain S141 cells at uniform levels of intensity from either the lac or thenpt-2 promoter, whereas expression of gfp could be visualized in Psychrobacter sp. strain SW5H cells at various levels of intensity only from the npt-2 promoter. Green fluorescent protein (GFP) fluorescence was not detected in the third species, Pseudoalteromonas sp. strain S91, when thegfp gene was expressed from either promoter. A new mini-Tn10-kan-gfp transposon was constructed to investigate further the possibilities of fluorescence tagging of marine bacteria. Insertion of mini-Tn10-kan-gfp generated random stable mutants at high frequencies with all three marine species. With this transposon, strongly and weakly expressed S91 promoters were isolated. Visualization of GFP by epifluorescence microscopy was markedly reduced when S91 (mini-Tn10-kan-gfp) cells were grown in rich medium compared to that when cells were grown in minimal medium. Mini-Tn10-kan-gfp was used to create an S91 chitinase-negative, GFP-positive mutant. Expression of the chi-gfp fusion was induced in cells exposed toN′-acetylglucosamine or attached to chitin particles. By laser scanning confocal microscopy, biofilms consisting of microcolonies of chi-negative, GFP+ S91 cells were found to be localized several microns from a natural chitin substratum. Tagging bacterial strains with GFP enables visualization of, as well as monitoring of gene expression in, living single cells in situ and in real time.
The gene encoding green fluorescent protein (GFP) has recently become an important visual marker of gene expression in eukaryotic organisms, as it is more sensitive than other reporter genes, requires no special cofactors for detection (7), and can be quantitated with a spectrofluorimeter (24). GFP has not been as widely applied to prokaryotic organisms because of a lack of constructs useful for diverse groups of bacteria, although GFP vectors are available for specialized bacterial systems (13, 24, 33, 41, 42). The wild-type gfpgene has been mutated to improve detection and expression of the fluorescent protein in prokaryotes (10, 18, 30), and both the wild-type and mutated forms have been used to construct less specialized bacterial GFP vectors.
A broad-host-range plasmid expressing the improved gfp(mut2) (10) gene from either a lac or annpt-2 promoter has been used successfully to tag gram-negative soil bacteria with GFP (27). Escherichia coli-Pseudomonas spp. shuttle vectors containinggfp(mut2) expressed from lac and tacpromoters have been constructed (5), and a GFP cloning cassette containing a similarly improved gfp gene is available for creating transcriptional fusions in prokaryotes (30). A suicide plasmid containing a promoterlessgfp that recombines wholly with the bacterial chromosome has been constructed to create genomic gfp fusions in a diverse range of gram-negative bacteria (22). Transposons provide an alternative method to insert reporter genes directly into the genomic DNAs of target strains. Several Tn5-based transposons containing either a promoterless gfp gene or agfp gene expressed from a broad-host-range promoter have been generated for use in tagging diverse bacterial species (6, 9, 27, 40). Tn5 reporter gene systems, however, are not effective in all gram-negative bacteria (2, 36).
GFP-tagged bacteria have been used in ecological studies to monitor single cells or cell populations in activated sludge communities (14) during symbiosis with plant cells (15), during infection of macrophages (24), during plasmid conjugation on semisolid surfaces (9), and in survival studies of E. coli in aquatic environments (26). There have been recent advances in detecting the presence of specific genes in single cells, thereby enabling identification of specific cells in mixtures by in situ PCR (20, 21), as well as in detecting the expression of genes in single cells by in situ PCR after an initial reverse transcription step targeting the mRNA in the cell (8, 39). Although these methods are useful approaches for many experiments, they all involve killing the cells because of the fixation step in the procedure and the heating steps in the PCR regimen.
We are interested in studying regulation of gene expression (37, 38), surface colonization behavior (11), and plasmid transfer (3) in several diverse marine bacterial species in situ. To investigate these processes in living microbial communities, it is necessary to find methods of tagging the different marine species so that cells may be visualized in situ and in real time. For this purpose we compared levels of expression of gfp(mut2) from two different bacterial promoters on a broad-host-range plasmid (27) in three gram-negative marine bacterial species. As Tn5-based transposons do not work with the marine bacteria we have tested and Tn10-based transposons do (2), we also constructed a Tn10-based reporter transposon containing promoterless gfp(mut2) (hereafter referred to asgfp) to create transcriptional fusions in order to investigate gene expression in marine bacteria further.
Here we show the differences in levels of expression of gfpin three vector-transposon constructs in three marine species:Pseudoalteromonas sp. strain S91, Vibrio sp. strain S141, and Psychrobacter sp. strain SW5H. GFP-tagged S91 cells were used to investigate initial biofilm formation on a natural biodegradable substratum, squid pen, and laser scanning confocal microscopy (LSCM) was used to visualize the hydrated structure of the biofilm at the squid pen surface.
MATERIALS AND METHODS
Bacteria, plasmids, and growth conditions.The bacterial strains and plasmids used in this study are described in Table1. The gfp reporter transposon constructed in this study is shown in Fig.1. The plasmid construct pLOFKmgfp in E. coli SM10 has been lodged with the American Type Culture Collection. E. coli strains were grown in Luria-Bertani broth (LB) (28) at 37°C. All other strains were grown at 30°C in either tryptone soy broth (Oxoid) containing NaCl (0.26 M), MgCl2 (1 mM), and CaCl2 (0.33 mM) (TS); LB containing NaCl (0.26 M), MgCl2 (1 mM), and CaCl2 (0.33 mM); or artificial seawater minimal medium (32) supplemented with 20 mM glutamate (MMMglt) for strains S91 and SW5H or 20 mM glucose for strain S141. Agar plates contained 15 g of Bitek agar (Sigma) liter−1 unless otherwise indicated. The following antibiotics (Sigma) and concentrations were used: ampicillin (50 μg ml−1), kanamycin (600 μg ml−1 for S91 and 100 μg ml−1 for all other strains), and streptomycin (100 μg ml−1).
Strains and plasmids used in this study
Diagrammatic representation of mini-Tn10-gfp-kan. Tn10inverted-repeat ends are shown as filled boxes at either end. Genes and relevant restriction enzyme sites are indicated; large arrows show the direction of gene transcription. Primers used in construction are shown above the boxes, with small arrows indicating the 5′-to-3′ direction. The diagram is not to scale.
DNA manipulations.Plasmid extractions, restriction enzyme digestions, ligations, transformations, and agarose gel electrophoresis were carried out by standard methods (35) and according to the manufacturers’ instructions where appropriate. Restriction and other enzymes were obtained from New England Biolabs Inc.
PCR.PCR amplification of the 740-bp gfp fragment from pBCgfp was done as previously described (27) with the following primers: gfpSfiI-F (5′-CTCCTCGGCCGCCTAGGCCGATTTCTAGATTTAAGAAGG) andgfpSfiI-R (5′-CTCCTCGGCCTAGGCGGCCTCATTATTTGTATAGTTCATC). PCR amplification of the 1.3-kb fragment from pLOFKmgfptransformants was carried out as described above withgfpSfiI-F and Kmseq-F (5′-TACAATCGATAGATTGTCGC), a primer designed to amplify sequence upstream from the kanamycin resistance (Kmr) gene of pLOFKm.
Conjugations.Mobilization of p519gfp, p519ngfp, and pLOFKmgfp from E. colihosts with E. coli(pNJ5000) as a helper was done by plate matings as previously described (2). The numbers of transconjugants, donors, and recipients from matings between E. coli SM10(pLOFKmgfp) and S91 were determined by plating a dilution series of each cell mix to MMMglt (kanamycin and streptomycin) and TS (kanamycin and streptomycin) to select transconjugants, LB (ampicillin) to select donors, and TS (streptomycin) to select recipients.
Screening for extracellular chitinase activity.Chitinase-negative mutants were screened on MMMglt supplemented with 0.1% yeast extract and 0.1% colloidal chitin as previously described (38).
Identification of the transposon-interrupted chitinase gene from S91CGFP.Part of the transposon-interrupted chitinase gene from S91CGFP was amplified by PCR amplification with a primer, PLOFOUT (5′-CACTGATGAATGTTCCGTTGC-3′), designed to extend outward from the 3′ end of the Kmr gene (38) and a primer, CHIAR1 (5′-ACCAATGTTGATGCGACC-3′), designed to extend inward from the 3′ end of a chitinase gene. The 500-bp PCR product obtained was used as a template for DNA sequencing with the PLOFOUT primer by the Taq–Dye-Terminator method on an automated DNA sequencer (Applied Biosystems model 373; DNA Sequence and Synthesis Facility, Westmead Hospital, Sydney, Australia).
Detection of GFP fluorescence and microscopy.Bacterial colonies on solid media were exposed to blue light in a light box constructed to contain a 100-W quartz-halogen lamp with an infrared filter and a 480-nm-band-pass filter (Andover Corp. part no., FS10-50).
An Olympus BX50 microscope, fitted with epifluorescence and differential interference contrast (DIC) optics, was used to visualize cells grown in liquid with and without colloidal chitin particles. Images were generated by either DIC or epifluorescence (excitation, 488 nm; emission, 520 nm) optics with a 40× oil objective lens, numerical aperture of 1.0. Images were recorded with a Panasonic digital closed-circuit television camera (model WV-BP510/A) and captured and prepared with NIH Image (version 1.59) and Adobe Photoshop (version 3.0.4) software, respectively, running on a model 7600/120 Power Macintosh. For photomicrography, slides were coated with gelatin (3%) to prevent cell movement.
For experiments involving LSCM, squid pen, which consists of 40% chitin and 60% protein (wt/wt) (16) was collected from a fish market in Sydney and stored at −80°C as described previously (38). Biofilms of S91CGFP cells were grown on 1-cm2 pieces of squid pen suspended in MMMglt at 30°C. After 24 h and then 7 days, small slices were cut aseptically from a piece of squid pen and placed without further treatment on a glass slide and covered by a coverslip, with MMMglt as the mounting medium.
A Bio-Rad MRC-1000 LSCM system in combination with a Nikon Diaphot 300 inverted microscope was used to obtain LSCM images of bacterial microcolonies attached to squid pen. The microscope was equipped with a 40×, 1.15-numerical-aperture water immersion lens and a krypton-argon laser. Excitation at 488/10 nm was used for GFP and chitin. Due to autofluorescence of squid pen at an emission of >515 nm, both GFP and the squid pen surface could be imaged. At an emission of 522/35 nm, only GFP was visualized. Images of microcolonies attached to squid pen were collected as xy and xz sections and captured as digital computer files, and quantitative examination was performed with CoMOS (Bio-Rad) computer image analysis software.
RESULTS AND DISCUSSION
Expression of GFP from a lac or an npt-2 promoter.p519gfp and p519ngfp are broad-host-range mob+ plasmids, derived from the broad-host-range RSF1010 derivative pDSK519 (23), that were constructed to contain gfp expressed from a lacor an npt-2 promoter, respectively (27). Thenpt-2 promoter is known to be more effective thanlac in gram-negative bacteria other than E. coli(4, 25, 27, 31). E. coli DH5α carrying either p519gfp or p519ngfp was conjugated separately to each of the three marine strains with the E. coli(pNJ5000) helper. Transconjugants were selected on TS (kanamycin and streptomycin) plates. Each marine strain carrying either p519gfp or p519ngfp was grown to exponential phase and assessed by epifluorescence microscopy for expression ofgfp. More than 99% of S141 cells expressed GFP uniformly at high intensity from either promoter. Fluorescence was so strong that colonies of S141 carrying either promoter could be easily identified on TS plates by eye. Although more than 99% of SW5H(p519ngfp) cells expressed GFP, fluorescence was not uniform. Of cells expressing GFP, only 14% (42 of 296) did so at high intensity. Variation in fluorescence of SW5H cells may have resulted from plasmid instability in this strain. pDSK519 was maintained poorly in SW5H cells, whereas the plasmid was well maintained in S141 and S91 cells (data not shown). Cells expressing GFP were not detected in SW5H(p519gfp) cultures, suggesting that the lac promoter was not functional in this strain. GFP fluorescence of S91(p519gfp) or S91(p519ngfp) cells during exponential growth could not be detected. Weak GFP fluorescence was detected, however, in S91(p519ngfp) cells after 2 days of growth as colonies on TS plates. It is possible that the npt-2 promoter functioned poorly in S91 such that a relatively longer time was required for GFP to accumulate to levels sufficient for visibility by epifluorescence microscopy but that the lac promoter was not functional at all. Bloemberg et al. reported that gfp was expressed poorly from a tac promoter in Pseudomonas aeruginosa (at a level 10 times lower than in E. coli) andPseudomonas fluorescens (at a level 20 times lower than inE. coli) and was not expressed from a lacpromoter in P. fluorescens (5).
Construction of pLOFKmgfp(mini-Tn10-gfp-kan).As gfp expressed from either promoter on pDSK519 derivatives was not useful for all three marine species tested, each species was tagged with gfp by transposon delivery direct to the chromosome. It is known that mini-Tn10 (19) yields stable transconjugants in our marine strains (2) but that various mini-Tn5, including mini-Tn5-gfp(27), or Tn5 transposons yield no transconjugants (reference 2 and data not shown). It was necessary therefore, to construct a mini-Tn10 with gfp as the reporter gene for use with the marine bacteria. The plasmid pLOFKm contains the mini-Tn10 transposon which carries the Kmr gene (19). A promoterless gfpgene was inserted in a position similar to that of the promoterlesslacZ gene in mini-Tn10-lac-kan carried by plasmid pLBT, which was described previously (2). A promoterlessgfp fragment, including the T7 (gene 10) ribosome binding site, was amplified from pBCgfp with primers containing anSfiI restriction enzyme site at their 5′ ends. The 740-bp product was digested with SfiI and ligated to pLOFKm, also digested with the same enzyme. The ligation mixture was transformed into E. coli SM10 competent cells, and transformants were selected for ampicillin resistance (Apr) and Kmr. Plasmid DNA was extracted from each transformant and linearized with SfiI, and fragments were separated by gel electrophoresis. One transformant that contained a correctly sized 740-bp insert was selected. Plasmid DNA from this transformant was amplified with the primers gfpSfiI-F and Kmseq-F. A 1.3-kb PCR product from this plasmid confirmed the correct orientation ofgfp with respect to pLOFKm. The plasmid was named pLOFKmgfp, and the transposon, mini-Tn10-gfp-kan (Fig. 1), was used to create transcriptional gfp fusions.
Use of pLOFKmgfp with marine bacterial strains.The mini-Tn10-gfp-kan transposon, delivered from pLOFKmgfp, transposed at high frequency in all three marine strains. Representative data for the recipient, S91, are given here. Of 1.7 × 108 CFU of E. coli donors ml−1 and 1.1 × 1010 CFU of S91 recipients ml−1, 2.0 × 105 CFU and 2.4 × 105 CFU of S91 Kmrtransconjugants ml−1 were recovered on TS (streptomycin and kanamycin) and MMMglt (streptomycin and kanamycin) plates, respectively. Transconjugants were patched to MMMglt (streptomycin and ampicillin) to test for loss or retention of the delivery vector (pLOFKmgfp). Of 192 Kmr transconjugants, 190 were ampicillin sensitive (Aps), indicating loss of the delivery vector in 99.5% of transconjugants. The mini-Tn10-gfp-kan transposon was assumed to give single random insertions in the three marine bacterial strains, as was shown for mini-Tn10-lac-kan (2). Strong S91 promoters, as well as chitinase-activity-negative S91 mutants, were selected and investigated further.
Expression of GFP from a strong S91 promoter.S91(mini-Tn10-gfp-kan) transconjugants were patched to MMMglt and monitored for fluorescence. Detection of GFP in S91 transconjugant colonies was not possible by eye under normal light. Of 250 transconjugant colonies, 25 (10%) could have been scored as fluorescent when they were exposed to blue light. As S91 has a distinct orange pigment, it was not possible, however, to differentiategfp fluorescent mutant colonies from pigment-negative mutant colonies by eye under blue light. Cells from 96 individual transconjugants were screened by epifluorescence microscopy to identify strongly expressed constitutive promoters. Of these, 32 (33%) were GFP positive, including 9 (9%) which showed strong fluorescence and were scored as strongly expressed promoter-gfp fusions. Interestingly, with all S91 transconjugants tested (several hundred), visualization of GFP by epifluorescence microscopy was markedly reduced when cells were grown on TS compared to that when they were grown on MMMglt, either on plates or in broth. It is not known whether this was due to a difference in levels of GFP expression in cells grown in either medium or whether it was due to production of an exterior cell polymer in S91 grown on TS that interfered with visualization of GFP fluorescence.
Use of GFP to investigate chi gene expression in living S91 cells.Approximately 2,000 S91 Kmr transconjugants were screened for altered chitinase activity by patching them onto MMMglt containing 0.1% colloidal chitin. Three mutants were presumptively determined to be chitinase negative after 7 days. One of these, S91CGFP, was identified as being gfp positive by epifluorescence microscopy. Transcription of gfp in this mutant was under the control of the promoter of an interruptedchi gene. DNA sequence information obtained from the chitinase-interrupted gene of S91CGFP revealed that the transposon mini-Tn10-gfp-kan had inserted into the same chi gene as that described for S91CX (38).
Using lacZ as a reporter gene, we previously quantitated induction and repression of this chi gene promoter in a population of S91CX cells in response to various environmental conditions and nutrient regimens (38). In order to use thischi-gfp transcriptional fusion mutant to investigate genetic regulation of chitinase genes in living single cells and in situ, S91CGFP was grown in MMMglt and MMMglt with either 0.1%N-acetylglucosamine or 0.1% colloidal chitin. Expression of the chi promoter was monitored by visualization ofgfp expression by epifluorescence microscopy. Thechi-gfp fusion was up-expressed when cells were exposed to either N-acetylglucosamine or colloidal chitin particles, which correlated with the quantitative data previously obtained withlacZ as a reporter gene (38) (Fig.2). S91CGFP cells attached to colloidal chitin particles were visible, although not all attached cells were fluorescent. Such variation in levels of expression of thechi gene promoter at cell surfaces highlights the difference between monitoring expression of environmentally regulated genes in single cells in real time and monitoring gene expression in a population of cells. With lacZ as a reporter gene, differential levels of expression of the algC promoter in single living P. aeruginosa cells attached to a surface were observed (12).
Microscope digital image demonstrating GFP fluorescence from a chitinase-negative, GFP-positive mutant (S91CGFP) grown on MMMglt (A and B), MMMglt plus 0.1% N-acetylglucosamine (C and D), or MMMglt plus chitin particles (E and F). Images shown in panels A, C, and E were taken with DIC optics. Panels D, E, and F show the same images, respectively, under epifluorescence microscopy.
Although S91CGFP was unable to degrade purified colloidal chitin because of the insertion of the transposon into the chigene, cells were still able to grow on squid pen, presumably by digesting the proteinaceous portion of the pen with one or more of the several extracellular proteases produced by the cells (1). To investigate biofilm formation on squid pen, S91CGFP was grown in MMMglt containing pieces of squid pen so that biofilms began to develop from cells attached to and growing on the biodegradable surface (38). Examination of horizontal-optical-section (xy) LSCM images showed that after 24 h, the surface appeared to be patchily colonized by S91CGFP cells which were mostly localized in microcolonies (data not shown). Vertical-section (xz) LSCM images indicated that microcolonies appeared to be localized a few micrometers away from the surface (Fig.3). After 7 days, microcolonies had increased in size and still appeared to occur patchily over the pen surface. Localization of the microcolonies several micrometers from the autofluorescent pen surface was clear, as shown in Fig.4. Investigation of the nature of this nonfluorescent zone between squid pen surface and S91CGFP microcolonies is in progress. There was no obvious pitting of the pen surface, in contrast to reports of pitting of solid mineral surfaces by attached microbial consortia. It may be that microorganisms will display different surface colonization strategies and behaviors when forming biofilms on organic versus mineral or inert solid surfaces. When examining biofilm structures, Dalton et al. found that S91 cells colonized glass poorly as single cells after 24 h but colonized well after a longer time, during which cells appeared to show coordinated behavior (11). A biofilm structure in which the majority of the cell mass was localized several microns away from the surface was found for SW5 colonizing a hydrophilic glass surface (11).
Vertical-section (xz) LSCM image demonstrating microcolonies of the GFP+ chitinase-negative mutant (S91CGFP) on the surface of squid pen after 24 h, with the gap between microcolonies and the autofluorescent squid pen estimated to be about 6 μm. The image was first captured at an emission of 522/35 nm so that only GFP was visualized; then the same image was captured at an emission of >515 nm so that both GFP and autofluorescent squid pen were visualized; these two images were then used to generate the composite image shown. Scale bar, 50 μm; arrow, squid pen.
Vertical-section (xz) LSCM image demonstrating a microcolony of the GFP+ chitinase-negative mutant (S91CGFP) on the surface of squid pen after 7 days, with the gap between the microcolony and the autofluorescent squid pen estimated to be about 20 to 25 μm. Scale bar, 25 μm; arrow, squid pen.
The use of LSCM in combination with GFP-tagged marine bacteria makes it possible to investigate the colonization behavior of bacteria, as pure and mixed species, on natural biodegradable substrata such as squid pen. It may now be possible to visualize, in real time, the biodegradation of particulate squid pen by GFP-tagged S91 wild-type cells and compare this to biodegradation by S91 GFP mutants containing various known genetic lesions in the chitinase system. Work is under way to construct a derivative of p519gfp in whichgfp expression is driven by the S91 chi promoter. This plasmid can then be introduced into the S91 parent strain, which will remain the wild type with respect to its chitinase activity, and this introduction will also enable visualization of the expression of the chi promoter in single living cells in real time.
No single promoter-vector system, or even culture medium, could be guaranteed to allow detection of GFP in bacteria isolated from natural environments. Conditions may need to be optimized to enable visualization of GFP in single cells of different bacterial species. This work demonstrated that mini-Tn10-kan-gfp is a useful tool for identifying promoters that are strongly or weakly expressed in marine bacteria. Such promoters can now be used for future fluorescence tagging of marine bacteria. Transposon mini-Tn10-kan-gfp was also useful in constructing transcriptional gfp fusions in marine bacteria, enabling visualization of gene expression in single living cells in situ and in real time.
ACKNOWLEDGMENTS
This work was supported in part by The Flinders University of South Australia. S. Stretton was supported by a Flinders University postgraduate scholarship, and S. Techkarnjanaruk was supported by a Royal Thai Government scholarship.
We thank Corbett Research for the thermal cycler, Doug Butler for constructing the blue-light box, and Peter Kolesik (Confocal Facility, Department of Horticulture, Waite Institute, The University of Adelaide) for expert assistance with the LSCM.
FOOTNOTES
- Received 22 December 1997.
- Accepted 20 April 1998.
- Copyright © 1998 American Society for Microbiology
