Present address: Guido V. Bloemberg, Institute of Medical Microbiology, University of Zurich, Gloriastr. 32, Postfach, CH-8006 Zurich, Switzerland.
Editor: Jeff Cole
Correspondence: Sandra de Weert, Institute Biology Leiden (IBL), Leiden University, Sylviusweg 72, 2333 BE Leiden, The Netherlands. Tel.: +31 71 527 5075; fax: +31 71 527 4900; e-mail: firstname.lastname@example.org
Live-cell imaging techniques are essential to gain a better understanding of microbial functioning in natural systems, for example in biofilms. Autofluorescent proteins, such as the green fluorescent protein (GFP) and the red fluorescent protein (DsRed), are valuable tools for studying microbial communities in their natural environment. Because of the functional limitations of DsRed such as slow maturation and low photostability, new and improved variants were created such as mCherry. In this study, we developed genetic tools for labeling Gram-negative bacteria in order to visualize them in vitro and in their natural environment without the necessity of antibiotic pressure for maintenance. mcherry was cloned into two broad host-range cloning vectors and a pBK-miniTn7 transposon under the constitutive expression of the tac promoter. The applicability of the different constructs was shown in Escherichia coli, various Pseudomonas spp. and Edwardsiella tarda. The expression of mcherry was qualitatively analyzed by fluorescence microscopy and quantified by fluorometry. The suitability of the constructs for visualizing microbial communities was shown for biofilms formed on glass and tomato roots. In addition, it is shown that mCherry in combination with GFP is a suitable marker for studying mixed microbial communities.
Live cell techniques are essential to gain a better understanding of microbial organization and functioning in vitro and in nature. The use of autofluorescent proteins for noninvasive microscopy is nowadays a well-established and valuable tool in biology and biotechnology. For studying microbial communities, multiple autofluorescent proteins can be applied simultaneously for visualization of different populations and intracellular processes. The use of red fluorescent protein (DsRed) in combination with enhanced green fluorescent protein (eGFP) is very suitable as the excitation and emission spectra of these proteins are well separated (Matz et al., 1999). In comparison with GFP, the use of DsRed has been hampered due to its longer maturation time (caused by its tetrameric form) and lower photostability. mRFP1 was the first monomeric derivative of DsRed, which has a shorter maturation time (Bevis & Glick, 2002). Subsequently, improved variants were developed with a more complete maturation and an over 10-fold increased photostability, of which mCherry is considered as one of the best alternatives for mRFP1 (Shaner et al., 2004). Tagging bacteria with marker genes is predominantly based on transformation of plasmids carrying the gene, which require antibiotic pressure for maintenance in the cell. Plasmids are attractive genetic tools for bacterial tagging due to their multicopy number, selective properties and easy handling for cloning strategies. In many natural environments, antibiotics cannot be applied for the efficient maintenance of plasmids (e.g. biofilms). However, cloning vectors that can be maintained without antibiotic selection are scarce. Alternatively, transposons can be used for stable integration in the chromosome, but have the disadvantage of being present as one copy per cell, which will result in a lower production of marker protein(s) in comparison with plasmids when using the same promoter.
Most bacteria form biofilms in their natural habitat (Costerton et al., 1995). Biofilms are defined as bacterial cells attached to a biotic or an abiotic surface, which are encased in an extracellular matrix (glycocalyx) mainly consisting of exopolysacharides. Studying biofilms is important because biofilm formation is commonly involved in bacterial infections, and plays an important role in industrial and agricultural processes. For example, Pseudomonas spp. that form biofilms on plant roots can protect plants against microbial diseases (Bloemberg & Lugtenberg, 2001). Microorganisms in a biofilm were shown to be more resistant to biocides, antibiotics and host immune responses (Costerton et al., 1999), which hampers the application of antibiotics for plasmid maintenance.
The aim of this work is to develop a set of genetic tools for tagging Gram-negative bacteria with mcherry that is constitutively expressed, can be maintained in the cell without antibiotic selection and is expressed at a level that allows visualization of single cells.
Materials and methods
Bacterial strains, culture conditions and plasmids
The bacterial strains and plasmids used in this study are listed in Table 1. Pseudomonas strains were grown at 28 °C in King B broth (King et al., 1954) or in a modified M63 minimal media (Pardee et al., 1959), for which M63 was supplemented with 1 mM MgSO4, 0.2% glucose and 0.5% casamino-acids. Antibiotics were added when required in the following final concentrations: tetracyclin, 40 μg mL−1; gentamycin, 10 μg mL−1; kanamycin, 50 μg mL−1; or streptomycin, 10 μg mL−1. Escherichia coli was grown in Luria–Bertani (LB) broth (Sambrook & Russel, 2001) at 37 °C. When appropriate, LB was supplemented with antibiotics in the following final concentrations: tetracyclin, 16 μg mL−1; carbenicillin, 100 μg mL−1; or kanamycin, 50 μg mL−1. Edwardsiella tarda was grown at 28 °C in tryptic soy broth (Becton Dickinson and Company, Sparks, MD). When required, the medium was supplemented with gentamycin (30 μg mL−1) or tetracyclin (16 μg mL−1).
Table 1. Bacterial strains and plasmids
Bacterial strains and plasmids
Reference or source
Wild type; excellent colonizer of grass roots. Production of lipopeptides putisolvin I and II
pMP6031 derivate harboring mCherry gene under the control of the tac promoter
pBBR1MCS-5 derivate harboring mCherry gene under the control of the tac promoter
pBK-miniTn7-KmΩSm1 derivate harboring mCherry gene under the control of the tac promoter
Growth curves were obtained by diluting an overnight culture to an OD620 nm of 0.1 in 20 mL of LB medium. Subsequently, cultures were grown for 24 h at 28 °C at 150 r.p.m.
Construction of plasmids
The mcherry gene was amplified with primer oMP1197 (5′-AAAAGGATCCGGGGAATTCTTGACAATTAATCATCGGCTCGTATAATGTGTGGAATTGTGAGCGGATAACAATTTTCACACAGGAAACAGCTAAATGGTGAGCAAGGGCGAG-3′), including a BamHI site (underlined) and the tac promoter (italics) and primer oMP1198 (5′-AAAGGATCCAAAACCGCCCTGCAAGGCGGTTTTTTCGTTTTCTTACTTGTACAGCTCGTCC-3′), including a BamHI site (underlined) and cloned into pGEM®-T Easy Vector System II (Promega Benelux, Leiden, the Netherlands), resulting in pGEM-mcherry.
From this construct, a BamHI fragment or a NotI fragment including mcherry and the tac promoter were cloned into plasmids pME6031 (Heeb et al., 2000), pBBR1MCS-5 (Kovach et al., 1995) and pBK-miniTn7 (Koch et al., 2001) (Fig. 1), resulting in plasmids pMP7604, pMP7605 and pMP7607, respectively (Fig. 1). Plasmids are publically available and will be supplied on request by the first author.
Transformation of plasmids and the stability of mCherry constructs
Bacterial strains were transformed with plasmids by conjugation according to standard methods (Sambrook & Russel, 2001) Conjugation of plasmids pMP7604 and pMP7605 was accomplished by mixing the donor E. coli DH5α containing pMP7604 or pMP7605, the helper E. coli strain containing pRK2013 and the recipient strains either Pseudomonas putida PCL1445, Pseudomonas fluorescens WCS365, Pseudomonas aeruginosa PAO1 or E. tarda FL60-60. Plasmid pMP7607 was introduced into P. putida PCL1445 for transposition via quadripartite mating using E. coli DH5α containing pMP7607, E. coli DH5α containing helper plasmid pRK2013 and E. coli DH5α containing pUX-BF13.
The stability of the mcherry containing constructs was analyzed by daily subculturing tagged strains (1 : 100) in liquid medium without antibiotics for approximately 30 generations. Each day, dilutions of the cultures were plated on LB plates without antibiotics. After colony formation, colonies were counted and analyzed for expression of mcherry using a Leica MZFLIII stereo fluorescence microscope (Leica, Wetzlar, Germany) (excitation 510/20 nm with 560/40 nm emission). This experiment was performed in triplicate and repeated once.
Quantification of mCherry expression
The production of mCherry in transformed strains was quantified using an HTS 7000 Bio Assay Reader (Perkin Elmer, Waltham, MA). Two hundred microliters of overnight cultures was transferred to a black 96-well flat-bottomed plate (Packard BioScience BV, Groningen, the Netherlands). Fluorescence was quantified by excitation at 590 nm with three flashes and by measuring the emission at 635 nm for 40 μs. The cell density of the cultures was determined by measuring a 1 : 10 dilution of the overnight culture at OD620 nm.
Fluorescence microscopy and confocal laser scanning microscopy (CLSM)
Planktonic cells of mcherry-labeled strains were studied using an Axioplan 2 microscope (Zeiss, Mannheim, Germany), equipped with filterset XF108-2, for which 10 μL of an overnight culture was used. Images were captured using an AxioCam MRc5 camera (Zeiss).
Bacteria attached to tomato roots and glass surfaces were visualized using an Axioplan epifluorescence microscope (Zeiss) coupled to an MRC 1024ES confocal system (Biorad, Hemel Hempstead, UK). Images were obtained using a Krypton/Argon laser using excitation 488 nm-emission 522/35 nm for eGFP and excitation 568–585 nm long pass emission for mCherry. The projections of the individual channels were merged using imagej 1.38 (Wayne Rasband, National Institutes of Health).
Biofilm formation on glass
Biofilm formation on glass was established by placing a microscopy glass slide in a 50-mL falcon tube containing 20 mL M63 medium to which 5 μL of an overnight culture was added. Tubes were incubated under nonshaking conditions at 28 °C for 24 h. A biofilm was formed in the middle of the glass slide at the liquid–air interface. Before microscopic analysis, the slide was rinsed carefully and a cover slip was placed on top. The biofilm was analyzed using CLSM as described above. To establish mixed biofilms, cultures of strains tagged with mCherry and eGFP were mixed in a 1 : 1 ratio.
Tomato root colonization assays
Root colonization assays were performed using the gnotobiotic system as described by (Simons et al., 1996). Coated tomato seedlings (a 1 : 1 ratio of bacterial strains) were placed in the gnotobiotic quartz sand system, moistened with a plant nutrient solution without a carbon source but with NO3 as a nitrogen source. After growth for 7 days, plants were removed from the system and were carefully washed with a phosphate-buffered saline solution. Roots were subsequently analyzed for the presence of bacterial biofilms using CLSM as described above.
Construction of Gram-negative strains expressing mCherry
To express mcherry in Gram-negative bacteria, the gene was cloned in two broad host-range vectors, i.e. pBBR1MCS-5 (Gmr) and pME6031 (Tcr) and in the miniTn7 transposon (Kmr) located on pBK-miniTn7 (Fig. 1). Plasmid pRSET-B-mCherry was used as a template for obtaining a PCR fragment of mcherry using primers oMP1197 (containing the tac promoter) and oMP1198 (Table 1). This resulted in a 785-bp PCR product, which was cloned into pGEM®-T EasyII and subsequently cloned into pME6031, pBBR1MCS-5 and pBK-miniTn7, resulting in pMP7604, pMP7605 and pMP7607, respectively (Fig. 1;Table 1). These plasmids were introduced into P. putida PCL1445, P. aeruginosa PAO1, P. fluorescens WCS365 and E. tarda FL6-60, which resulted in bright red fluorescent colonies as observed by fluorescence microscopy. One colony from each transformation or transposition event was selected for the following studies.
Growth rate and stability
Growth in liquid LB medium of P. putida PCL1445 transformed with pMP7604, pMP7605 and pMP7607 and their corresponding empty vectors was followed. The expression of mcherry as well as the presence of the vectors had no significant effect on growth compared with the wild-type strain P. putida PCL1445 (data not shown).
The stability of the plasmids and the transposon integration was tested by subculturing in nonselective media (without antibiotic selection pressure) for approximately 30 generations. Samples of the subcultures were plated and colonies were screened for the expression of mcherry by fluorescence microscopy. Strain PCL1481 carrying miniTn7∷mcherry did not show any loss of integration. No loss of plasmid was observed for PCL1479 carrying pMP7604, whereas 3% of the colonies of strain PCL1480 carrying pMP7605 had lost fluorescence at day 3 (data not shown).
Qualitative and quantitative analysis of mCherry expressed in P. putida PCL1445
A qualitative and quantitative analysis for mCherry production in P. putida PCL1445 tagged with pMP7604, pMP7605 and pMP7607 was performed in order to evaluate the resulting brightness of the different constructs. Cells of overnight cultures were visualized using fluorescence and light microscopy (Fig. 3a) and fluorescence was quantified using fluorometry (Fig. 2b). mcherry expression was detected at the single-cell level for all tagged strains. Microscopic and fluorometric analyses showed that strain PCL1480 (harboring pMP7605) produced the highest amount of mCherry and strain PCL1481 (containing miniTn7-mcherry) produced the lowest amount (Fig. 3a and b). The strains PCL1479, PCL1480 and PCL1481 produced mCherry in a ratio of 15 : 95 : 1, respectively. No significant fluorescence was detected for P. putida PCL1445 cells and strains PCL1477 and PCL1478 containing the cloning vectors pME6031 and pBBR1MCS-5 (Fig. 2b).
Expression of mCherry in Gram-negative bacteria
To evaluate the applicability of the mCherry marker vectors for tagging Gram-negative bacteria, several other Gram-negative spp., such as P. fluorescens WCS365 (an efficient root colonizer), P. aeruginosa PAO1 (a model strain for cystic fibrosis research) and E. tarda FL6-60 (a fish pathogen and model for zebrafish immunology), were transformed with pMP7604 and pMP7605. This yielded PCL1700, PCL1701, PCA0241, PCA0242, PCA0239 and PCA0240, respectively. Fluorescence microscopy analysis showed the production of mCherry for all transformed strains (data not shown). Single colonies were isolated and overnight cultures were grown for quantitative analysis of mCherry production and comparison with P. putida PCL1445 (Fig. 4). Strains containing pMP7605 showed the highest mCherry production. Comparable mCherry production levels were observed among the four strains tested, except for the one carrying pMP7605, which showed a lower level of expression in E. tarda FL6-60.
Visualization of mCherry-tagged P. putida PCL1445 strains in biofilms formed on glass and tomato roots
To analyze the applicability of the mcherry-expressing constructs pMP7604, pMP7605 and pMP7607 in established test systems, which are not suitable for efficient application of antibiotic pressure, P. putida PCL1445-tagged strains were allowed to form biofilms on glass (in vitro biofilm assay) and on tomato roots (in vivo assay used to study root colonization). Using CLSM, the tagged strains were visualized at the single-cell level in both assays (Fig. 4a–f). Biofilms formed on glass consisted of a homogenous spread layer. In contrast, biofilms on tomato roots as formed for 7 days of growth after seedling inoculation were visualized as distinct colonies formed at the interjunctions between the root cells. The brightest fluorescence signal was produced by P. putida PCL1480 cells, followed by PCL1479 and PCL1481, which is consistent with the quantitative fluorometric data of these strains (Figs 2 and 3).
Simultaneous visualization of bacterial populations tagged with mCherry or eGFP in biofilms
In order to analyze the use of the mcherry-expressing constructs in combination with egfp for simultaneous visualization, differentially tagged bacterial populations of the same strain were allowed to form biofilms and were subsequently visualized by CLSM (Fig. 5). Because the egfp is cloned in a similar vector as pME6031 and is also expressed under control of the Ptac promoter, pMP7604 was selected for testing simultaneous visualization. CLSM analysis of the biofilms formed on glass (Fig. 5a and b) showed clearly the presence and distinction between mcherry- and egfp-tagged bacteria. For tomato root colonization experiments, P. putida PCL1445 strains harboring pMP7604 (PCL1479) or pMP7605 (PCL1480) (Fig. 5d) were used for mixed inoculation (1 : 1) of seedlings with P. putida PCL1445 tagged with egfp. CLSM analysis of the roots after 7 days of growth clearly showed the presence of mixed microcolonies originating from the mcherry- and egfp-tagged populations (Fig. 5c and d).
Nowadays, the use of autofluorescent proteins as markers for the noninvasive microscopic analysis of biological processes is a well-established successful technical approach (Errampalli et al., 1999; Larrainzar et al., 2005; Bloemberg, 2007). Autofluorescent proteins with sufficiently separated excitation and emission spectra are required for simultaneous visualization of (1) interactions between different bacterial populations or various spp. and (2) metabolic processes. GFP has been extensively optimized for codon usage in different organisms (Patterson et al., 1997) and its intrinsic characteristics such as photostability, brightness and excitation/emission spectrum (Shaner et al., 2007). GFP is the most frequently used marker gene in biology and biotechnology. Excitation and emission spectra of GFP and red fluorescent protein (Matz et al., 1999) hardly overlap, which makes their combination suitable for simultaneous application (Tecon et al., 2009). In order to improve brightness, maturation and photostability optimized monomeric forms of red fluorescent protein have been produced recently, of which mCherry is one of the best members (Shaner et al., 2004, 2005). mCherry has been used successfully in several recent studies, as a reporter, and also as a biosensor (Hillson et al., 2007; Lewenza et al., 2008; Malone et al., 2009).
We have cloned mcherry under the control of the tac promoter, which is expressed constitutively at a low level, into the vectors pBBRMCS-5 (Kovach et al., 1995) and pME6031 (Heeb et al., 2000) and into the transposon vector pBK-miniTn7 (Koch et al., 2001) (Fig. 1). The performance of these genetic tools for tagging various Gram-negative bacteria was compared. The three different vectors were chosen for their difference in antibiotic selection gene (gentamycin, tetracyclin and kanamycin, respectively) and the opportunities for maintenance as a plasmid (pBBRMCS-5 and pME6031) or integration into the chromosome (pBK-miniTn7). In addition, pBBRMCS-5 (a derivative of the general cloning vector pBBR) is assumed to have a higher copy number than pME6031 (containing the pVS1 replicon). pME6031 was described as being maintainable without the selective pressure of tetracyclin (Heeb et al., 2000). All vectors were reported to have a broad host range in Gram-negative bacteria.
Pseudomonas putida strain PCL1445, which is an excellent root colonizer and is able to form biofilms on abiotic surfaces such as polyvinylchloride (Kuiper et al., 2004a), was selected to examine the new constructs containing mcherry. Growth curves of the transformed strains did not show an effect of the constructs and mcherry expression on growth (data not shown). However, care should be taken when using these plasmids under other growth conditions. As expected, the pME6031-derived plasmid pMP7604 was maintained without antibiotic pressure (no loss was observed), whereas the pBBRMCS-5-derived plasmid pMP7607 showed a loss of 3% in cells of the population after 3 days of subculturing without antibiotic pressure. Qualitative and quantitative analyses showed that all constructs can be used for visualization at the single-cell level and that the intensity of fluorescence resulting from the use of the different genetic constructs correlates with the copy number of the different plasmids and the transposon used (Fig. 2). The mcherry constructs created were shown to be functional in different Pseudomonas spp. (i.e. P. putida PCL1445, P. fluorescens WCS365 and P. aeruginosa PAO1) and the fish pathogen E. tarda, with comparable mCherry production levels (Fig. 3). In addition, fluorescence was observed during cloning in E. coli.
Labeled strains under in vitro (biofilm formation on glass) and in vivo (tomato root colonization) conditions showed that the constructs are well suited for the visualization at the single-cell level (Figs 4 and 5). In addition, tagging with the mcherry plasmid constructs was shown to be useful for the simultaneous visualization with the eGFP-tagged strain of P. putida PCL1445 as shown for biofilms formed on glass and tomato roots (Fig. 5). Also, single strains tagged with eGFP and mCherry were recently shown to be useful for bioreporter studies (Tecon et al., 2009). The vectors constructed in this study could function as markers to locate bacteria in such studies.
In conclusion, we have developed a set of genetic tools for the expression of mcherry in Gram-negative bacteria for studies in vitro and in natural environments, especially of value when antibiotic pressure cannot be efficiently applied.
We would like to thank Joost van Soest and Merle Eijkhof for their technical assistance. We are grateful to the Tsien lab (University of California, San Diego) for obtaining pRSET-B-mCherry and Ole Nybroe for providing pBK-miniTn7.
S.d.W. and G.V.B. contributed equally to this work.