A Mycobacterium marinum promoter, designated G13, was isolated from a promoter-trap library as a constitutive producer of the mutant green fluorescent protein. Sequence analysis, primer extension analysis, and computer promoter prediction analysis indicate that the G13 promoter is very similar to Escherichia coli consensus σ70 promoters. Expression of the green fluorescent protein from the G13 promoter in M. marinum is, however, up to 40 times higher than that seen from the mycobacterial hsp60 promoter during exponential growth. Further, expression from this promoter does not appear to affect the growth of the organism in culture media or in macrophages. The strong expression of the G13 promoter allows it to be developed as a useful molecular tool for high level expression of markers in vitro.
Mycobacterium marinum is a pathogen classified as a slow-growing member of the genus Mycobacterium[1, 2]. Human infection with M. marinum results in cutaneous lesions known as ‘swimming pool granuloma’, and in severely immunocompromised patients can cause fatal systemic disease . Phylogenetic analyses such as DNA/DNA homology and 16S RNA analysis indicate that M. marinum is closely related to Mycobacterium tuberculosis. M. marinum is a facultative intracellular pathogen which replicates and persists in mammalian cells . Like M. tuberculosis, M. marinum resides in macrophage vesicles which do not undergo phagolysosomal fusion or acidification . Although classified as a slow grower, M. marinum has a doubling time of 4–6 h (compared to 20 h for M. tuberculosis) and may be manipulated in the laboratory at biosafety level 2. Genotypic and phenotypic similarities between M. marinum and M. tuberculosis, the more rapid growth of M. marinum cultures, and the ease of laboratory and genetic manipulations of M. marinum make this organism a good model system for the study of mycobacterial gene expression and pathogenesis.
There have been relatively few detailed studies of mycobacterial promoters (for review, see ). A recent study by Bashyam and Tyagi  described an extended −10 promoter in which transcription of mycobacterial promoters was enhanced by a TGN motif directly upstream of mycobacterial promoters. The TGN sequence was found in approximately 20% of the promoters studied. The same researchers found that mycobacterial promoter sequences were very different from Escherichia coliσ70 consensus sequences with very few of the mycobacterial promoters functional in E. coli. The Mycobacterium bovis P2hsp60 promoter (hsp60), which contains a TGN motif in the extended −10 region, has been relatively well characterized and is considered a strong mycobacterial promoter . In this study, we describe and characterize a novel M. marinum promoter with a σ70-like consensus sequence that drives high expression of green fluorescent protein (GFP) in M. marinum. We further show evidence that this promoter is equally effective in Mycobacterium smegmatis and E. coli.
Materials and methods
Bacterial strains and culture conditions
M. marinum 1218R (American Type Culture Collection strain ATCC 927), M. smegmatis strain mc2155 (provided by William R. Jacobs, Albert Einstein College of Medicine, New York), and E. coli DH5α were grown and maintained as previously described  in media supplemented with 30 µg ml−1 kanamycin as needed.
Macrophage and media growth and fluorescence assays
For fluorescence and growth assays, mycobacteria were grown in 7H9 medium supplemented with 10% Middlebrook oleic acid, albumin, dextrose, and catalase enrichment (OADC), in tissue culture media, i.e., Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), or in RAW 264.7 (ATCC TIB71) murine macrophages. The macrophage cell line was maintained in DMEM supplemented with 10% FBS at 37°C in 5% CO2. Macrophages were infected with bacteria at a multiplicity of infection (MOI) of five bacteria per macrophage. At each time point, bacteria were harvested from macrophage infections using methods described previously . Macrophage lysates and bacterial cultures were diluted and plated for viability counts, and analyzed by FACS as described previously .
Plasmid and strain designations
Promoter-trap plasmid pFPV27 was constructed and provided by Raphael Valdivia, Stanford University, CA. Plasmid pMV262gfpmut3, constructed and provided by Lalita Ramakrishnan, Stanford University, CA, contains the mycobacterial hsp60 promoter driving expression of the mGFP gene. Plasmid pG13 contains the G13 promoter sequence ligated into pFPV27 upstream of the mGFP gene. Transformants of M. marinum strain 1218R containing pMV262gfpmut3 or pG13 were designated GFP3R and G13R, respectively . Plasmid pG13 was also transformed into M. smegmatis mc2155 and E. coli strain DH5α. These transformants were designated mc2155-G13 and DH5α-G13, respectively.
Plasmid mini-preps were performed on G13R as previously described , and PCR amplification of inserts was performed according to previously described protocols . All DNA sequencing was performed using a model 370A Stretch automated DNA sequencer according to the manufacturer's instructions (Applied Biosystems, Foster City, CA). Forward and reverse primers for use in PCR reactions with M. marinum chromosomal DNA as the template were 5′-TGCAGAACTTTCATGAATTAGGCC-3′ and 5′-GCCATTATCGCGGCTATGACT-3′, respectively. The PCR product was sequenced as described above.
RNA isolation and primer extension
Total RNA was isolated from M. marinum G13R, M. smegmatis mc2155-G13, and E. coli DH5α-G13 in the following manner. Cultures were grown to late exponential phase as determined by absorbance at 600 nm. Cells were pelleted by centrifugation (14 000×g, 5 min, 4°C) and resuspended in approximately 1 ml of Ultraspec (Biotec Labs Inc., Houston, TX). Approximately 0.3 g of glass beads (Biospec, Bartlesville, OK) were added to the bacteria/Ultraspec solution followed by shaking at high speed for 5 min with a Biospec bead beater. Samples were centrifuged (2000×g, 5 min, 4°C) and the supernatant was removed. RNA was isolated from the supernatant as described by the manufacturer (Biotec Labs Inc.). RNA resuspended in 100 µl of DEPC treated water was quantitated by UV absorbance and checked for integrity and DNA contamination by formaldehyde gel as previously described by Porcella et al. .
Primer extension analysis was performed using the Promega (Madison, WI) AMV Reverse Transcriptase Primer Extension System as described by the manufacturer. Two primers were designed for primer extension and designated PE-1 (5′-CCGCAACACTTCTGCCATTATCGC-3′) and PE-2 (5′-TGGGACAACTCCAGTGAAAAGTTCTTCTCC-3′). The locations of PE-1 and PE-2 are shown in Fig. 2. DNA sequencing with both primers was used in parallel to identify the transcription initiation site using the Perkin Elmer (Foster City, CA) AmpliCycle Sequencing kit and following the manufacturer's protocols. Sequencing and primer extension products for each primer reaction were resuspended in loading buffer and were loaded on a 6% polyacrylamide gel. Gels were run at 50 W continuous power for 3 h, transferred to blot paper, and exposed to autoradiographic film.
The G13 promoter
The isolation of the G13 promoter was described previously [11, 16]. To ascertain that the cloned mycobacterial DNA fragment was indeed responsible for the high expression of mGFP, the plasmid containing the G13 insert was isolated, designated pG13, and transformed into wild-type 1218R. This strain, designated G13R, exhibited the same high levels of mGFP expression as the original clone.
Promoter strength and growth studies
The G13R strain was assayed for fluorescence and growth rate and compared to wild-type M. marinum 1218R and strain GFP3R carrying the hsp60-GFP plasmid. The results of the growth determination and fluorescence experiments are shown in Fig. 1 and Table 1, respectively. The exponential growth rate of G13R did not vary appreciably from that of GFP3R and the wild-type 1218R in 7H9 or DMEM media, nor in macrophages. In highly saturated 7H9 or DMEM stationary phase cultures, however, it did appear that the G13R strain was maintained at a slightly lower cell density than that of GFP3R or 1218R. The average fluorescence of individual G13R bacteria was consistently higher in media and in macrophages than the fluorescence of the GFP3R bacteria. In actively growing media cultures, the G13R strain exhibited 5–40-fold higher fluorescence than the GFP3R strain. In stationary phase cultures, where the G13R strain accumulates high levels of mGFP, there was up to 120-fold more fluorescence in G13R as compared with GFP3R. Because M. marinum is cytotoxic to the host cell , growth and fluorescence in macrophages were measured only up to 6 days. In the macrophage, growth rates were comparable between all three strains (see Fig. 1), but G13R fluorescence was 2–10 times higher than the GFP3R strain (Table 1).
DNA sequence analysis
The sequence of the G13 insert is shown in Fig. 2. The G13 insert was determined to be 457 bp long, and homologies to known sequences were determined with BLAST software (National Center for Biotechnology Information at the National Library of Medicine) . The G13 insert was homologous (73% identity) to a non-coding sequence of M. tuberculosis DNA upstream of an uncharacterized M. tuberculosis open reading frame. A potential promoter within the G13 insert was identified by computer analysis using the program MacTargsearch. This algorithm is based upon σ70-like consensus sequences identified in E. coli.
Primers internal to the G13 insert sequence were designed for use in a PCR reaction with M. marinum chromosomal DNA as the template. The sequence of the fragment from the M. marinum chromosome was 100% identical to the cloned G13 sequence shown in Fig. 2.
RNA isolation and primer extension analysis
Results from the primer extension analysis performed with primer PE-1 are shown in Fig. 3. A single primer extension product identical in size for M. smegmatis mc2155-G13 and M. marinum G13R was produced and, when compared to the adjacent sequencing marker lanes, allowed for the determination of the potential start site (Fig. 3). Primer extension analysis performed on RNA isolated from the E. coli DH5α-G13 recombinant using primer PE-1 produced a fragment identical in length to that seen for G13R and mc2155-G13 (data not shown), suggesting that E. coli uses the same promoter and initiation site as M. marinum and M. smegmatis. Primer extension analysis with primer PE-2 for all three bacteria supported the potential initiation site produced by PE-1, albeit at lower resolution, but failed to show any other discernable products which might indicate alternative initiation sites downstream or upstream of the one identified by primer PE-1 (data not shown). These primer extension analysis data indicate that a single, identical initiation site exists for the G13 fragment in E. coli, M. marinum and M. smegmatis.
We have described and characterized a mycobacterial promoter expressed at very high levels in media and in the macrophage. Indeed, this is the most highly expressed promoter identified to date for mycobacterial species. The strong expression of the G13 promoter allows for the development of useful molecular tools for high expression of markers in vitro and in vivo. We have previously shown that pG13 expresses mGFP at levels higher than the pMV262gfpmut3 construct when transformed into M. smegmatis, M. tuberculosis, and M. bovis. Although it has been shown previously that most mycobacterial promoters function poorly in E. coli, the G13 construct expresses mGFP in E. coli at levels high enough to see fluorescent colonies when the E. coli is grown on solid media. Even though expression of the mGFP from the G13 promoter is much higher than from the hsp60 promoter in all backgrounds, very high levels of G13 promoter-driven mGFP production do not appear to impair the growth of M. marinum in media nor in macrophages.
A comparison of the G13 promoter with other promoter sequences [8, 20–22] in Fig. 4 demonstrates that the G13 promoter shows σ70 consensus-like homology. The previously identified mycobacterial extended −10 promoters, which includes the hsp60 promoter, are divergent from the G13 promoter. Further, the conserved TGN motif described in the terminal spacer region of the extended −10 promoters is absent from the same spacer region for the G13 promoter.
It has been reported that mycobacterial promoters lack a consensus −35 region and are expressed poorly in E. coli[8, 19]. Conversely, mycobacterial promoters described that are homologous to the E. coliσ70 consensus usually do not express well in mycobacterial backgrounds. This study is the first to demonstrate a mycobacterial σ70-like promoter which functions strongly in both E. coli and Mycobacteria species.
The sequences of the −10 and −35 regions of M. marinum pG13 are identical to the corresponding sequences in M. tuberculosis, except for a ‘C’ to ‘G’ difference in the −10 position. Studies are under way to determine whether other transcriptional factors or secondary structures are involved in the transcriptional regulation of the G13 promoter.
It is likely that the G13 promoter will serve as a very useful tool for the over-expression of mycobacterial genes and markers in infections with M. marinum and other pathogenic mycobacteria. In addition, the G13 promoter has the potential for identifying weakly expressed virulence determinants expressed in the macrophage environment that might otherwise be overlooked. Lastly, the G13 promoter may be useful in vaccine studies where over-expression of target molecules would be useful.
Special thanks to James Carroll, Patti Rosa, and Kathleen M. George for their critical review of the manuscript. Thanks also to Gary Hettrick and Bob Evans for help with graphics and to Lalita Ramakrishnan and Raphael Valdivia for the provision and helpful discussion of vectors.