ABSTRACT
Lyme disease is the most widely reported vector-borne disease in the United States. Its incidence is rapidly increasing and disease symptoms can be debilitating. The need to understand the biology of the disease agent, the spirochete Borrelia burgdorferi, is thus evermore pressing. Despite important advances in B. burgdorferi genetics, the array of molecular tools available for use in this organism remains limited, especially for cell biological studies. Here, we adapt a palette of bright and mostly monomeric fluorescent proteins for versatile use and multi-colour imaging in B. burgdorferi. We also characterize two novel antibiotic selection markers and establish the feasibility of their use in conjunction with extant markers. Lastly, we describe a set of constitutively active promoters of low and intermediate strengths that allow fine-tuning of gene expression levels. These molecular tools complement and expand current experimental capabilities in B. burgdorferi, which will facilitate future investigation of this important human pathogen.
INTRODUCTION
Lyme disease, a widespread infection transmitted by hard ticks of the Ixodes genus, is the most prevalent vector-borne disease in the United States. The condition is also common in Europe and Asia, and its incidence and geographic distribution have been steadily increasing in recent decades (1). Lyme disease is caused by spirochetal bacteria belonging to the Borrelia burgdorferi sensu lato group, with B. burgdorferi sensu stricto (from here-on referred to as B. burgdorferi) being the principal agent in North America, and B. afzelii and B. garinii being the primary agents in Eurasia. In humans, acute Lyme disease is often associated with a characteristic skin rash and flu-like symptoms. If left untreated, late stages of infection may result in carditis, neurological manifestations, and arthritis (2).
Spirochetes in general, and the Borrelia species in particular, display cellular features unusual for bacteria (3). Spirochete cells are typically very long and thin by bacterial standards. B. burgdorferi cells, for example, are 10 to 25 μm long and ~ 250 nm wide (4-6). Spirochetes are also highly motile, but, unlike in most bacteria, their flagella are not external organelles (7). Instead, their flagella are located in the periplasm (i.e., between the inner and outer membranes). In B. burgdorferi, the helicity of the flagella imparts the flat-wave morphology of the bacterium (8). B. burgdorferi also possesses what is likely the most segmented genome of all bacteria analysed to date. It is made up of a linear chromosome of about 900 kilobases (kb) and over 20 linear and circular genetic elements ranging from 5 to 60 kb in length (9,10). These smaller genetic elements are often referred to as plasmids, though many of them encode proteins that are essential for the life cycle of this organism (11). Recent work from our laboratory has shown that Borreliae species also have an uncommon pattern of cell wall synthesis in which discrete zones of cell elongation in one generation predetermine the division sites of daughter cells in the next generation (6).
While these unusual cellular features are integral to B. burgdorferi physiology and pathogenesis, little is known about how they arise or are maintained over generations. In fact, the cell biology of this pathogen remains largely unexplored. Technical hurdles have slowed progress in this area. Genetic manipulation of B. burgdorferi is feasible, but the available genetic tools are still limited, and the process remains cumbersome (12,13). Constitutive gene expression is mostly limited to the use of very strong promoters. Moreover, apart from a few exceptions (14-19), fluorescent protein reporters have primarily been used as gene expression reporters or as cellular stains for in vivo localization of the spirochete (13). Yet fluorescent proteins have many more uses, which have transformed the field of cell biology (20). For example, fluorescent proteins have opened the door to localization studies in live cells. They have also facilitated the detection of protein-protein interactions, the measurement of physical properties of the cell, and the analysis of single events and of population heterogeneity. Much of this information is not accessible through the use of bulk biochemical measurements on cell populations.
The averaging inherent to such techniques leads to loss of spatial resolution and obscures rare events and cell-to-cell or subcellular heterogeneity of behaviour (21). Indeed, the ability to perform extensive genetic manipulations and to use a wide panel of fluorescent proteins in an organism has been key to progress in understanding bacterial cell biology (22). Such approaches have been extensively used in model bacteria such as Bacillus subtilis, Escherichia coli, and Caulobacter crescentus since the first reported use of fluorescent protein fusions two decades ago (23-25). In order to facilitate the study of B. burgdorferi, we have generated new investigative tools by characterizing a panel of fluorescent proteins, promoters and antibiotic resistance markers for use in this medically important bacterium.
MATERIAL AND METHODS
Bacteria, growth conditions, and genetic transformations
Bacterial strains used in this study are listed in Table 1. E. coli strains were grown at 30 °C in liquid culture in Super Broth medium (35 g/L bacto-tryptone, 20 g/L yeast extract, 5 g/L NaCl, 6 mM NaOH) with shaking, or on LB agar plates. Plasmids were transformed by electroporation or heat shock. For selection of E. coli strains we used 200 μg/mL (solid medium) or 100 μg/mL (liquid medium) ampicillin, 20 μg/mL (solid medium) or 15 μg/mL (liquid medium) gentamicin, 50 μg/mL kanamycin (solid and liquid media), 50 μg/mL spectinomycin (solid medium), 50 μg/mL streptomycin (liquid medium), and 25 μg/mL (liquid medium) or 50 μg/mL (solid medium) rifampicin.
B. burgdorferi strains were grown in BSK-II medium supplemented with 6% (vol/vol) heat inactivated rabbit serum (Sigma Aldrich or Gibco) or in complete BSK-H medium (Sigma Aldrich), as previously described (26-28). Cultures were incubated at 34 °C under 5% CO2 atmosphere in a humidified incubator. Antibiotics were used at the following concentrations (unless otherwise indicated): gentamicin at 40 μg/mL, streptomycin at 100 μg/mL, kanamycin at 200 μg/mL, blasticidin S at 10 μg/mL, and hygromycin B at 250 μg/mL. Ampicillin was purchased from Fisher Scientific, blasticidin S and hygromycin B from Invivogen, and all other antibiotics and biliverdin hydrochloride from Sigma Aldrich.
B. burgdorferi strain generation
B. burgdorferi electrocompetent cells were prepared as previously described (29) and were transformed with shuttle vector plasmid DNA (usually 30 μg) by electroporation. Electroporated cells were then allowed to recover overnight in BSK-II medium at 34 °C. The next day, antibiotic selection was initiated and 5-fold serial dilutions of the culture were plated in a 96-well plate (24 wells for each dilution). After 10-14 days of incubation, the wells were inspected by microscopy using dark-field illumination. When fewer than 20% of the wells of a given dilution were positive for growth, those wells were considered to contain clonal populations and were further expanded and characterized. When appropriate, fluorescence imaging was used to confirm fluorescent protein expression. Alternatively, selected, non-clonal transformant populations were enumerated using C-Chip disposable hemacytometers (INCYTO), using the manufacturer’s instructions with the following change: counting was done by continuously scanning the full height of the counting chamber for each counting surface to account for the height of the counting chamber being larger than the size of the spirochetes. Enumerated spirochetes were then diluted in BSK-II media and plated in 96-well plates at an average density of 0.2 cells/well. After 10-14 days, clonal growth was confirmed by dark-field microscopy imaging.
Determination of minimal inhibitory concentrations (MIC) and antibiotic cross-resistance
MICs were determined using strains B31 e2 or B31 MI, while cross-resistance testing was done using B31 e2-derived strains that contained shuttle vectors carrying kanamycin, gentamicin, streptomycin, blasticidin S, or hygromycin B resistance markers (see strains CJW_Bb069 through CJW_Bb073 in Table 1). For both tests, antibiotics were two-fold serially diluted in complete medium. For each concentration, 100 μL of antibiotic solution were dispensed into two to four wells of 96-well plates. The cell density of B. burgdorferi cultures was determined by direct counting using dark-field microscopy. The cultures were then diluted to 2 × 104 cells/mL in antibiotic-free medium, and 100 μL of this diluted culture were added to the antibiotic-containing wells to yield an inoculum of 104 cells/mL. The plates were incubated for at least 4 days at 34 °C under 5% CO2 atmosphere in a humidified incubator, after which each well was checked for spirochete growth and motility using dark-field microscopy. A well was marked as positive if motile cells were detected. The plates were further incubated for several days, during which bacterial growth-dependent acidification caused the phenol-red pH indicator in the medium to change colour. This colour change was documented using colourimetric trans-illumination imaging on a GE Amersham Imager 600. We verified that growth scoring of each well by dark-field imaging matched the observed medium colour change.
DNA manipulations
Plasmids used in this study are listed in Table 2. Methods of plasmid construction and sequences of oligonucleotide primers are provided as Supplementary Data. Standard molecular biology techniques were used, as detailed in the Supplementary Data. Codon optimisation was performed using the web-based Java Codon Adaptation Tool hosted at www.jcat.de (30), using the codon usage table for B. burgdorferi as stored at www.kazusa.or.jp/codon (31). Codon-optimised DNA sequences were then chemically synthesized at Genewiz. DNA sequences of each codon-optimised gene are provided in the Supplementary Data. The names of these genes include a Bb superscript to indicate that the gene’s nucleotide sequence is codon-optimised for translation in B. burgdorferi (e.g. iRFPBb). The name of the protein encoded by such a gene (e.g. iRFP), however, does not include the Bb superscript, as the protein’s amino acid sequence does not differ from that expressed from other versions of the gene.
Microscopy
Visualization and counting of live spirochetes were done using a Nikon Eclipse E600 microscope equipped with dark-field illumination optics and a Nikon 40X 0.55 numerical aperture (NA) phase contrast air objective. Phase contrast and fluorescence imaging was done on a Nikon Eclipse Ti microscope equipped with a 100X Plan Apo 1.40 NA phase contrast oil objective, a Hamamatsu Orca-Flash4.0 V2 Digital CMOS camera, a Sola light engine (Lumencor), and controlled by the Metamorph software (Molecular Devices). Alternatively, light microscopy was performed on a Nikon Ti microscope equipped with a 100X Plan Apo 1.45 NA phase contrast oil objective, a Hamamatsu Orca-Flash4.0 V2 CMOS camera, a Spectra X light engine (Lumencor) and controlled by the Nikon Elements software. Excitation of iRFP was achieved using the 640/30 nm band of the SpectraX system, but higher excitation efficiency and thus brightness could in theory be obtained using a red-shifted excitation source between 660 and 680 nm. The following Chroma filter sets were used to acquire fluorescence images: CFP (excitation ET436/20x, dichroic T455lp, emission ET480/40m), GFP (excitation ET470/40x, dichroic T495lpxr, emission ET525/50m), YFP (excitation ET500/20x, dichroic T515lp, emission ET535/30m), mCherry/TexasRed (excitation ET560/40x, dichroic T585lpxr, emission ET630/75m), and Cy5.5 (excitation ET650/45x, dichroic T685lpxr, emission ET720/60m). For imaging, cultures were inoculated at densities between 103 and 105 cells/mL, and were grown for two to three days to reach densities between 106 and 3×107 cells/mL. The cells were then immobilized on a 2% agarose pad (6,32) made with phosphate buffered saline covered with a No. 1.5 coverslip, after which the cells were immediately imaged live. Images were processed using the Metamorph software. Figures were generated using Adobe Illustrator software.
Image analysis
Cell outlines were generated using phase contrast images and the open-source image analysis software Oufti (33). Outlines were checked visually for each cell and were extended manually to the full length of the cells when appropriate. When not assigned to single cells or assigned to non-cellular debri outlines were manually removed. The remaining outlines were automatically refined using the Refine All function of the software. Fluorescence signal data was added to the cells in Oufti. The resulting cell lists were processed using the MATLAB script addMeshtoCellList.m (see Supplementary Data for the code). This script uses the functions CL_getframe.m, CL_removeCell.m, CL_cellId2PositionInFrame.m and getextradata.m which were previously described (33). Single cell fluorescence intensity values were calculated by dividing the total fluorescence signal inside a cell mesh by the cell mesh area using the MATLAB-based function CalculateFluorPerCell.m. Final fluorescence data were plotted using the GraphPad Prism 5 software. The number of cells analysed for each condition is provided in the Supplementary Data.
RESULTS
A wide palette of fluorescent proteins for imaging in B. burgdorferi
Only a few fluorescent proteins have been used to date in B. burgdorferi (summarized in Table 3). These proteins belong primarily to two colour classes: green fluorescent proteins (GFP) and red fluorescent proteins (RFP) (Table 3). To expand the range of options for multi-colour imaging of B. burgdorferi, we focused on a set of fluorescent proteins that have been used in localization studies in other organisms and codon-optimised their genes for translation in B. burgdorferi. The selected proteins span five colour classes (Table 3) and their signal can be collected using widely available filter sets for cyan fluorescent protein (CFP), GFP, yellow fluorescent protein (YFP), mCherry/TexasRed and Cy5.5 fluorescence. The selected cyan, green, and yellow variants are all derivatives of the jellyfish (Aequorea victoria) GFP. We used both the classic variants Cerulean (34), enhanced GFP, or EGFP (35), Citrine (36), as well as the superfolder (e.g., sfGFP) variants (37). All variants included the monomeric mutation A206K (38), denoted by a lower case m before the name of the protein (e.g., mCerulean). Our red protein of choice was mCherry (39), a monomeric, improved variant of mRFP1. Lastly, we codon-optimised and expressed an infrared fluorescent protein, iRFP (40). The far-red wavelengths used to excite this fluorophore are less toxic to cells than the shorter excitation wavelengths used for the other fluorescent proteins, and the sample autofluorescence in the near-infrared spectral region is lower than in the other, blue-shifted imaging windows (20,41).
To visualize these fluorescent proteins, we expressed them in strain B31 e2 from the strong flagellin promoter PflaB (42) located on a shuttle vector. With the exception of iRFP, each fluorescent protein displayed bright fluorescence when imaged using a filter set matched to its colour (Figure 1A). Unlike the other fluorescent proteins, which oxidatively conjugate their own amino acid side chains to create a fluorophore (20), iRFP covalently binds an exogenous biliverdin molecule, which then serves as the fluorophore (40). Adding the biliverdin cofactor to the growth medium of the iRFP-expressing strain rendered the cells fluorescent in the near-infrared region of the spectrum, as detected with a Cy5.5 filter set (Figure 1B). Treating a control strain carrying an empty shuttle vector with biliverdin did not cause any increase in cellular fluorescence (data not shown). To measure cellular fluorescence levels, we chose a microscopy-based approach in conjunction with quantitative image analysis. This allowed us to efficiently analyse hundreds of cells and to clearly distinguish individual cells from similarly-sized debris found in the culture medium, or from clumps of multiple cells. Using this method, we established that a 4 μM concentration of biliverdin in the growth medium was sufficient to achieve maximal cellular brightness (Figure 1C). Close-to-maximal iRFP brightness was reached as early as an hour after addition of biliverdin to the culture and was maintained throughout subsequent growth (Figure 1D). Furthermore, continuous growth of B. burgdorferi in the presence of biliverdin was indistinguishable from growth in biliverdin-free medium (Figure 1E). This indicates that culture experiments that involve iRFP may be performed either by adding biliverdin shortly before imaging or by growing the cells continuously in the presence of biliverdin.
In microscopy studies, simultaneous imaging of multiple fluorescent proteins requires that the signal generated by a given fluorescent protein does not overlap with the fluorescence channels used to collect the signal of another protein. To assess the viability of using our palette of fluorescent proteins for multi-colour imaging in B. burgdorferi, we quantified the signal generated by each fluorescent protein when imaged with the commonly used CFP, GFP, YFP, mCherry, and Cy5.5 filter cubes (Figure 2). We found that each fluorescent protein generated a strong signal when imaged with a colour-matched filter set (Figure 2). As expected, we detected a significant spectral overlap between CFP and GFP, as well as between GFP and YFP variants. Importantly, signal quantification showed that mCerulean or msfCFP can be imaged alongside mCitrine, mCherry, and iRFP, while mEGFP or msfGFP can be imaged alongside mCherry and iRFP, opening the door to combinatorial imaging of up to four proteins in the same B. burgdorferi cell.
Promoters for various levels of constitutive expression in B. burgdorferi
To date, constitutive expression of exogenous genes in B. burgdorferi, including antibiotic selection markers and reporter genes such as fluorescent proteins and luciferases, has almost exclusively relied on very strong promoters such as PflaB and PflaB (13,42). Reporter expression from strong promoters facilitates spirochete detection, particularly in high fluorescence background environments such as the tick midgut or mammalian tissues (43-45). However, as overexpression can affect protein localization, interfere with function, or cause cellular toxicity (e.g. (46-56)), low gene expression has proven instrumental in facilitating localization studies (e.g. (57-60)) and is often preferred in such applications.
To identify constitutively expressed promoters of low and medium strengths, we mined a published RNA sequencing (RNA-seq) dataset that measured transcript levels in cultures of B. burgdorferi in early-exponential, mid-exponential and stationary phases (61). We selected five genes whose expression was largely unchanged among the three growth phases tested (Figure 3A), amplified a DNA region upstream of each gene’s translational start site, and fused it to an mCherry reporter in a kanamycin-resistant shuttle vector (Figure 3B). The amplified putative promoter sequences ranged in size between 129 and 212 base pairs (bp) and included the reported 5’ untranslated regions (5’UTR) of these B. burgdorferi genes (61,62). We also included in our analysis an empty vector and a vector containing a PflaB-mCherryBb fusion, which served as references for no and high expression, respectively.
We transformed these constructs into B. burgdorferi, imaged the resulting strains, and quantified the fluorescence level in each cell. All promoters elicited fluorescence levels above the background of the strain carrying the empty vector (Figure 3C). We noticed differences between the RNA-seq and mCherry reporter methods of measuring promoter strength, as detailed in the discussion. Importantly, however, the promoters we tested displayed a broad dynamic range from low (P0526) to intermediate (P0826, PresT, P0031, and P0026) to high (PflaB) strength.
Antibiotic selection in B. burgdorferi using hygromycin B and blasticidin S resistance markers
Several antibiotic resistance markers have been used to perform genetic manipulations in B. burgdorferi and have recently been reviewed in detail (13). The most widely used today are the kanamycin (aphI), gentamicin (aacC1), streptomycin (aadA) and erythromycin (ermC) resistance genes (see Table 4) (42,63-65). Use of several other antibiotics for selection is either ineffective (e.g. zeocin, chloramphenicol, and puromycin), discouraged due to safety concerns (e.g. tetracyclines, β-lactams, and sometimes erythromycin), redundant due to cross-resistance (several aminoglycoside antibiotics), or no longer widespread (coumermycin A1) due to alterations in cell physiology induced by both the antibiotic and the resistance marker (13,64).
To expand the panel of antibiotic resistance markers that can be used in B. burgdorferi, we focused on two antibiotics commonly used for selection of eukaryotic cells, namely the translation inhibitors hygromycin B and blasticidin S. Rendering B. burgdorferi resistant to them does not pose a biosafety concern, as these antibiotics are not used to treat Lyme disease. We found that hygromycin B and blasticidin S prevented B. burgdorferi growth in liquid culture at concentrations of 200 and 5 μg/mL, respectively (Table 4). For resistance cassettes, we used the E. coli gene hph (also known as aph(4)-Ia), which encodes a hygromycin B phosphotransferase, and the Aspergillus terreus gene bsd, which encodes a blasticidin S deaminase (66-68). We codon-optimised these genes for translation in B. burgdorferi and placed them under the control of the strong PflgB promoter on a shuttle vector (Figure 4A). The resulting vectors, pBSV2H and pBSV2B, also carry the rifampicin resistance gene arr-2 of Pseudomonas aeruginosa (69-71), which encodes a rifampicin ADP-ribosyltransferase. B. burgdorferi is naturally resistant to rifampicin (72,73), but the use of rifampicin for selection in E. coli instead of the more expensive blasticidin S and hygromycin B antibiotics reduces the cost of generating and propagating the vectors in E. coli.
B. burgdorferi strains obtained by transforming pBSV2B or pBSV2H into B31 e2 grew readily in cultures containing 10 μg/mL blasticidin S or 250 μg/mL hygromycin B, respectively. We used these strains to test whether the antibiotic resistance cassettes encoded by these vectors conferred any cross-resistance to the often-used antibiotics kanamycin, gentamicin, streptomycin, and erythromycin. In parallel, we performed reciprocal tests using B31 e2-derived strains that carried a kanamycin, gentamicin, or streptomycin resistance cassette. Each culture was grown in the presence of two-fold serial dilutions of each antibiotic (Figure 4B). Each dilution series was centred on the concentration routinely used for selection with each of the tested antibiotics (Figure 4B, arrow). We grew the cultures for at least four days and then inspected each well for growth by dark-field imaging. A well was considered to be growth-positive if we detected at least one motile spirochete after scanning a minimum of five fields of view. In addition, we further incubated the plates to allow for growth-dependent acidification of the medium. This pH change is easily detected as a medium colour change from red, denoting no growth, to orange or yellow, denoting various degrees of growth (Figure 4C-H) (64). We confirmed that wells with the lowest antibiotic concentration at which the medium remained red also did not contain motile spirochetes. This concentration was taken to represent the minimum inhibitory concentration, or MIC (Figure 4C, black line). Whenever we exposed a strain to the antibiotic to which it carried a resistance gene, we readily detected growth at all antibiotic concentrations tested (Figure 4D-H), highlighting the efficacy of each resistance marker. Importantly, we did not detect any major cross-resistance between the five resistance markers and the six antibiotics tested (Figure 4D-H). One exception was the kanamycin-resistant strain CJW_Bb069, which was able to grow in the presence of as much as 40 μg/mL gentamicin (Figure 4E), a concentration routinely used for gentamicin selection (64). A slightly higher amount of gentamicin (80 μg/mL) was, however, sufficient to kill this kanamycin-resistant strain (Figure 4E). This low level of cross-resistance may thus necessitate use of a higher dose of gentamicin for selection if the parental strain is already kanamycin-resistant.
DISCUSSION
We have undertaken this work to facilitate microscopy-based investigations of the biology of the Lyme disease agent B. burgdorferi. We expanded the available molecular toolkit by characterizing antibiotic resistance markers, fluorescent proteins and constitutively active promoters not previously used in this organism.
Alongside the commonly used kanamycin, gentamicin, streptomycin, and erythromycin selection markers, the addition of hygromycin B and blasticidin S resistances as useful selection markers will provide more flexibility in designing genetic modifications. A wider array of non-cross-resistant selection markers is particularly important in the absence of a streamlined method to create unmarked genetic modifications in this bacterium (13). Currently, in infectious B. burgdorferi strains, an antibiotic resistance marker is commonly used to inactivate the restriction modification system encoded by the bbe02 locus on plasmid lp25. This inactivation increases the efficiency of transformation with shuttle vectors. It also helps maintain this plasmid in the cell population during in vitro growth through selective pressure (74-77). This step is essential for maintaining a strain’s infectivity, as linear plasmid lp25 is essential in vivo but is often rapidly lost during genetic manipulations and growth in culture (78,79). A second resistance marker is often used to inactivate a gene of interest, either by targeted deletion or by transposon insertion mutagenesis. A third resistance marker is needed for complementation, either at the original locus, or in trans. Additional markers are needed if two genes are to be inactivated and complemented simultaneously, or if several localization reporters need to be expressed both simultaneously and independently.
Today’s cell biology investigations often rely on microscopy studies using fluorescent protein fusions. Prior to our work, green and red fluorescent proteins have been the reporters of choice in B. burgdorferi microscopy studies (Table 3), and only a handful of subcellular localization and topology studies had been performed using these tools (13-19). We have expanded the palette of fluorescent proteins that can be used in this bacterium by adding several proteins with properties highly desirable for imaging and localization studies. These fluorescent proteins are among the brightest of their classes (20,37,40), and their spectral properties render them appropriate for simultaneous multi-colour imaging of up to four targets. For the most part, they are also monomeric, as all of the A. victoria GFP, CFP, and YFP variants that we have generated carry the A206K mutation (38). Using monomeric fluorescent proteins may be important to prevent artifactual intermolecular interactions, (e.g. (38,80,81)). Should the weakly dimeric versions of these proteins be required for specific applications, the A206K mutation can be easily reversed by site-directed mutagenesis. The superfolder variants of these proteins may facilitate tagging when the folding of the fusion protein is otherwise impaired (37). In addition, unlike EGFP, which does not fold in the periplasm of diderm bacteria when exported through the Sec protein translocation system, sfGFP does fold in the periplasm (82). It can therefore be an alternative to mRFP1 and mCherry for tagging periplasmic and outer-surface-exposed proteins. This is particularly relevant for the study of B. burgdorferi since this bacterium expresses an unusually large number of lipoproteins that are localized on the cell surface or in the periplasmic space (83). In addition, although dimeric, iRFP may serve as a useful in vivo marker, and may be preferable to GFP and RFP. Excitation light penetrance in live tissues is better in the far-red/near-infrared region of the spectrum than in the blue-shifted regions used to excite GFP or RFP. Furthermore, tissue autofluorescence in this spectral region is lower, which further facilitates imaging (84,85). Lastly, the levels of biliverdin found in animal tissues are in the low milimolar range, with healthy human plasma containing 0.9-6.5 μM biliverdin (86). In our hands, such biliverdin levels are sufficient to elicit maximal fluorescence of B. burgdorferi-expressed iRFP. Furthermore, iRFP has been successfully used to label Neisseria meningitidis bacteria for in vivo imaging (87). Altogether, these considerations indicate that imaging in mice using iRFP-expressing B. burgdorferi should be feasible.
We also characterized promoters of low and intermediate strengths and demonstrated that variable degrees of constitutive gene expression can be easily achieved in B. burgdorferi. The relative order of promoter strength, as quantified using the mCherry reporter (Figure 3E), largely matched the order of the expression levels of the corresponding genes in culture (Figure 3A) (61), with the exceptions of P0526 and P0826. While P0526 had an intermediate strength as measured by RNA-seq, it was the weakest when tested using our reporter system. In contrast, P0826 was the weakest promoter based on RNA-seq data, but displayed intermediate strength in our experiments. The differences may arise from strain differences or from our use of short DNA sequences of 129 to 212 bp, which presumably contain minimal promoter sequences. Any native regulatory elements located further upstream of these short promoter sequences are thus absent in our reporter plasmids. Differences in expression levels may also be caused by reporter expression from circular shuttle vectors. P0526 and P0826 are natively located on the chromosome and differences in DNA topology, including supercoiling, between the chromosome and the circular plasmids are known to affect gene expression in B. burgdorferi (88,89). Finally, it is worth noting that while both bb0526 and bb0826 encode leaderless transcripts, bb0826 has a secondary transcriptional start site located 54 bp upstream of the translational start site (62). This difference may also partly explain the observed promoter strength mismatch between the native gene and reporter fusion. Regardless of the reason for these discrepancies, these promoters will facilitate complementation and localization studies where medium and low gene expression levels may be required.
In summary, our study describes novel molecular tools that we hope will aid investigations in the Lyme disease field and spur further progress in the study of this medically important and highly unusual bacterium.
AVAILABILITY
Sequences of all the plasmids constructed in this study are available upon request. The DNA sequences of the various genes that were codon-optimised for expression in B. burgdorferi are provided in the Supplementary Data. The MATLAB code used to process cell fluorescence data is also provided as Supplementary Data.
ACCESSION NUMBERS
None.
SUPPLEMENTARY DATA
Supplementary Data are available in the accompanying document. It contains detailed plasmid construction methods, a list of oligonucleotide primer sequences used in this study, DNA sequences of genes that were codon-optimised for translation in B. burgdorferi, a record of cell numbers for each figure describing quantitative fluorescence data, MATLAB code used in this study, and supplementary references.
FUNDING
This work was supported by the Howard Hughes Medical Institute. C.J.-W. is an Investigator of the Howard Hughes Medical Institute.
CONFLICT OF INTEREST
The authors are aware of no conflict of interest.
ACKNOWLEDGEMENT
We thank Nicholas Jannetty for help with cloning experiments and Dr. Bradley Parry for help with the computational analyses. We are also grateful to Drs. Brandon Jutras and Patricia Rosa, as well as to the members of the Jacobs-Wagner laboratory for helpful discussions and/or critical reading of the manuscript.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.
- 16.
- 17.
- 18.
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.
- 25.↵
- 26.↵
- 27.
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.
- 45.↵
- 46.↵
- 47.
- 48.
- 49.
- 50.
- 51.
- 52.
- 53.
- 54.
- 55.
- 56.↵
- 57.↵
- 58.
- 59.
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.
- 68.↵
- 69.↵
- 70.
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.
- 76.
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.
- 92.
- 93.
- 94.
- 95.
- 96.
- 97.
- 98.
- 99.
- 100.
- 101.
- 102.
- 103.
- 104.