Codon-optimized DsRed fluorescent protein for use in Mycobacterium tuberculosis

Objective We have previously codon-optimized a number of red fluorescent proteins for use in Mycobacterium tuberculosis (mCherry, tdTomato, Turbo-635). We aimed to expand this repertoire to include DsRed, another widely used and flexible red fluorescent protein. Results We generated expression constructs with a full length DsRed under the control of one of three strong, constitutive promoters (Phsp60, PrpsA or PG13) for use in mycobacteria. We confirmed that full length DsRed (225 amino acids) was expressed and fluoresced brightly. In contrast to mCherry, truncated versions of DsRed lacking several amino acids at the N-terminus were not functional. Thus, we have expanded the repertoire of optimized fluorescent proteins for mycobacteria.


Introduction
Fluorescent proteins (FPs) have become the work horses of molecular biology and microbiology, with numerous applications. A plethora of variants of Aequorea victoria green fluorescent protein (GFP) [1] and Discosoma sp red fluorescent protein (DsRed) [2] are available covering almost the whole light spectrum from green to infrared [3]. Mutant derivatives have been engineered with altered excitation and emission wavelengths, increased or decreased stability, resistance to photo bleaching, sensitivity to environmental stimuli and substrates, as well as time for fluorophore maturation, intrinsic brightness and multimeric formats [3,4]. We previously described the use of a range of red reporters, of which the brightest was mCherry [5]. We wanted to expand our repertoire of proteins. Since DsRed has been widely used as a bright and stable reporter, we optimized constructs for its expression in M. tuberculosis.

BMC Research Notes
*Correspondence: Tanya.Parish@idri.org 2 Infectious Disease Research Institute, Seattle, WA 98102, USA Full list of author information is available at the end of the article GAG TTC together with the reverse primer DsRed-R 5′-AAG CTT TTA CAG GAA CAG GTG GTG CCG-3′. The restriction sites are underlined, potential start codons are in bold. The ORFs were excised and cloned into pSMT3 [6] as BamHI/HindIII fragments to generate pBlazeA1, pBlazeB1 and pBlazeC1 with DsRed under the control of the hsp60 promoter (Table 1). Plasmids pBlazeC8 and pBlazeC10 were generated by replacing P hsp60 with P rpsA and P G13 respectively. All three promoters should drive constitutive high level expression [5,7,8].

Quantitation of fluorescence in whole cells
Mycobacterium tuberculosis was electroporated as described [9] and transformants selected with hygromycin. M. tuberculosis was grown to stationary phase, harvested, washed twice in 10 mM Tris pH 8.0 and resuspended in 10 mM Tris pH 8.0 to an OD 580 of 0.25, 0.10, 0.05 and 0.01 in 12 × 100 mm glass culture tubes. Fluorescence was measured on a Shimadzu RF-1501 spectrofluorimeter (Shimadzu) with a detection range of 0-1015 relative fluorescent units at Ex/ Em 558/583 nm [5].

Western analysis of fluorescent proteins
Cell extracts were prepared from liquid cultures. Cells were harvested by centrifugation, washed twice in 10 mM Tris (pH 8.0), resuspended in 1 ml of 10 mM Tris (pH 8.0), and added to lysing matrix B tubes (QBiogene). Cells were disrupted using the Fastprep (QBiogene) set at speed 6.0 for 30 s. Samples were centrifuged at 4000 rpm for two min, and the supernatant was recovered and filter sterilized (0.2 micron filter). Protein was quantified using a BCA kit (Pierce), and 10 μg of total protein was subjected to Western blot using a rabbit anti-body (Clonetech). The primary antibody was detected using horseradish peroxidase goat-anti-rabbit (Sigma), and activity was detected using an ECL kit (GE Healthcare).

Results
We were interested in the use of FPs in M. tuberculosis and had previously used these as reporters of bacterial viability for in vitro and in vivo studies [5,8]. We were successful in obtaining high level expression by using codon-optimized versions of red fluorescent proteins driven by strong mycobacterial promoters [5].

Optimization of DsRed expression
We wanted to expand the range of reporters available for use to increase flexibility and allow dual reporter expression and monitoring. We selected DsRed for optimization, based on its Ex/Em wavelengths, and the fact that it is a well-characterized FP in wide use [3,4,[10][11][12][13][14].

Expression of DsRed uses a different translational start site than mCherry
Our initial attempts to obtain expression of a codonoptimized DsRed were unsuccessful. We constructed a synthetic gene for DsRed using a similar approach as we used with another red fluorescent protein mCherry (Fig. 1). We designed a codon-optimized version based on the DsRed-T3 protein previously used. We cloned the synthetic version into a mycobacterial expression vector and tested for fluorescence in M. tuberculosis. Surprisingly, we did not detect any fluorescence from this construct (Fig. 1c).
mCherry is a variant of DsRed and we expected the two proteins would be similarly functional. Our previous work demonstrated that mCherry is expressed from a distal translational start site than the one annotated in the databases [15]. Sequence alignment shows the few mutations which differ between the two (Fig. 2a); these include loss of the translational start site we identified for mCherry, although there are still multiple translational start sites (Fig. 1a). The version we used for the  synthetic gene used a downstream translation start site and would produce a truncated version of DsRed as compared to mCherry. Therefore it was possible that we did not express the full protein (Fig. 1b). In order to determine the functional start site for DsRed we used a different approach in which we cloned several versions of the  [19]. b Activity of full length DsRed expressed from mycobacterial promoters. pBlazeC1-P hsp60 ; pBlazeC8-P rpsA ; pBlazeC10-P G13 . Recombinant M. tuberculosis was resuspended in 10 mM Tris pH 8.0 to an OD 580 of 0.25, 0.10, 0.05 and 0.01 in 12 × 100 mm glass culture tubes. Fluorescence was measured at Ex/Em 558/583 nm. Data are the average ± SD of three cultures. c Plasmids were transformed into E. coli and cell-free extracts analyzed by Western blotting; 10 µg protein were subjected to SDS-PAGE, blotted onto PVDF membrane and visualized with anti-DsRed antibody. Lane 1-E. coli (no plasmid); Lane 2-pRed1; Lane 3-pRedA1; Lane 4-pRedB1; Lane 5-pRedC1; Lanes 6 and 7-pBlazeC1; Lane 8-pBlazeC8; Lane 9-pBlazeC10. The arrow indicates the size of the DsRed protein coding region into the expression vector under the control of the constitutive hsp60 promoter (Fig. 2b).
In order to test this, we used PCR amplification to extend the region sequentially. We extended the gene to incorporate both additional start sites and generate proteins of 214, 220 and 225 amino acids. These variants were cloned into the same mycobacterial expression system and tested. Plasmids were transformed into M. tuberculosis and fluorescence was monitored. In contrast to mCherry, expression of a functional fluorescent DsRed was not seen with any truncated versions of the protein.
In fact fluorescence could only be detected when the full length amino acid sequence (as annotated) was cloned into the expression vector; high level fluorescence was seen with transformants carrying the plasmid pBlazeC1 (Fig. 2b).
We constructed two alternative vectors with DsRed under the control of either P rpsA or P G13 (pBlazeC8 and pBlazeC10 respectively); both of these constructs gave high level expression in M. tuberculosis. Western blotting using an anti-DsRed antibody in E. coli demonstrated that a protein of the expected size was only seen in bacteria carrying the full length construct (pBlazeC series), but not in the strains carrying the truncated version ( Fig. 2c; lanes 5-9). Two additional bands are see in the Western, these are unknown proteins, but are also present in the control E. coli (no plasmid, Fig. 2c, lane 1).

Discussion
We have determined that the functional translational start sites for two closely related FPs are different in M. tuberculosis. Although mCherry was functional even when a truncated version was being expressed, DsRed was non-functional in a truncated form and only fluoresced when expressed as a full length protein (225 amino acids). Western blotting suggested that the lack of fluorescence was most likely due to a lack of protein expression, since proteins could not be detected in the plasmids carrying truncated forms. This difference may relate to protein stability, with the extended N-terminal portion of DsRed increasing stability or protein maturation; alternatively this could be attributed to the physical state of the active proteins, since mCherry functions as a monomer, whereas DsRed is a tetramer which might also affect protein degradation.
Fluorescent proteins have proved useful in multiple applications in mycobacteria; our previous constructs using mCherry have been widely disseminated to the community and used in a range of methods. For example, we have used these for high throughput drug testing [16], and imaging infection using animal models [8]. Other approaches have used mCherry to develop reporter strains for environmental sensing [17].
In conclusion, we have codon-optimized DsRed for use in M. tuberculosis and demonstrated its high level fluorescence in that species from three different promoters of slightly varying strength (hsp60, rpsA, and G13). These vectors extend our current repertoire of functional fluorescent proteins for mycobacteria. They will be useful for generating fluorescent strains of M. tuberculosis for use in multiple studies, such as monitoring drug efficacy in vitro and in vivo [5,8,16,18] and will allow for detection of multiple reporters simultaneously.