A versatile vector for mycobacterial protein production with a functional minimized acetamidase regulon

Abstract

Recombinant protein expression is a prerequisite for diverse investigations of proteins at the molecular level. For targets from Mycobacterium tuberculosis it is favorable to use M. smegmatis as an expression host, a species from the same genus. In the respective shuttle vectors, target gene expression is controlled by the complex tetra‐cistronic acetamidase regulon. As a result, the size of those vectors is large, rendering them of limited use, especially when the target proteins are expressed from multi‐cistronic operons. Therefore, in the current work we present a versatile new expression vector in which the acetamidase regulon has been minimized by deleting the two genes amiD and amiS. We assessed the functional properties of the resulting vector pMyCA and compared it with those of the existing vector pMyNT that contains the full‐length acetamidase regulon. We analyzed the growth features and protein expression patterns of M. smegmatis cultures transformed with both vectors. In addition, we created mCherry expression constructs to spectroscopically monitor the expression properties of both vectors. Our experiments showed that the minimized vector exhibited several advantages over the pMyNT vector. First, the overall yield of expressed protein is higher due to the higher yield of bacterial mass. Second, the heterologous expression was regulated more tightly, offering an expression tool for diverse target proteins. Third, it is suitable for large multi‐protein complexes that are expressed from multi‐cistronic operons. Additionally, our results propose a new understanding of the regulation mechanism of the acetamidase regulon with the potential to construct more optimized vectors in the future.

Introduction

Mycobacterial diseases, especially tuberculosis caused by Mycobacterium tuberculosis, continue to pose a serious threat to public health. Ongoing efforts to develop better diagnostic tools, drugs, and efficient vaccines demand recombinant proteins. For the heterologous expression of mycobacterial proteins, for example targets from the pathogen M. tuberculosis, several different hosts have been used, such as Escherichia coli, insect cells using baculovirus, Streptomyces, and M. smegmatis.1 The latter is most suited as an expression host due to its close relationship to M. tuberculosis, in particular due to the high G/C‐content of mycobacterial genomes (>65%)2 and similar codon usage.1, 3 To use M. smegmatis for the recombinant deoxyribonucleic acid (DNA) technology, a shuttle vector has to be designed. All published vectors for mycobacteria are based on the pMV260 series, which has been designed by combining two plasmids, pAL5000 from M. fortuitum and pUC19.4, 5 The resulting vector has two kinds of origins of replication, OriE for propagation in E. coli and OriM for the mycobacterial replication machinery. An additional important factor for the recombinant protein production is the controlled inducibility of the employed promoter. The promoters of l‐lactamase from M. fortuitum,6, 7 of tetracycline,7 and of acetamidase3, 6, 8, 9 have been used for protein expression using M. smegmatis as an expression host. Of these three, the acetamidase promoter system has been shown to have the most suitable properties for recombinant protein production.3, 9

Shuttle vectors for the heterologous overexpression of target genes using the acetamidase system were based upon the fact that M. smegmatis is capable of using acetamide as a sole carbon source.10 For this pathway, the bacteria have developed a relatively complex regulation cascade, ultimately producing the enzyme acetamidase (gene product of amiE). The 2.6 kb region lying upstream of amiE (Msmeg_5335) encodes for four open reading frames (ORFs): three regulatory genes amiA (Msmeg_5338), amiD (Msmeg_5337), and amiC (Msmeg_5339) plus amiS (Msmeg_5336) which encodes for a possible transporter.8, 11 Within this region, four respective promoters preceding each ORF have been identified (Fig. 1(A)): P_c drives the expression of amiC and has a constitutive low level activity; P₂, located upstream of amiD, shows increased activity in the presence of the inducer; P₁, situated upstream of amiA is constitutively active at a high level, and directly upstream of amiE there is the weak promoter P₃.8, 12 To take into account this complexity we propose the term “acetamidase regulon” instead of acetamidase “promoter” or “operon” currently used in the literature.6, 9, 13 The P₂ promoter region is the site of negative regulation by the gene product of amiA, which is due to direct DNA binding.13 AmiC is thought to have an acetamide binding site. Upon binding of the AmiC‐inducer complex to AmiA, the P₂ promoter is released which results in the expression of the genes amiD, amiS and amiE.11, 12 The production of AmiE (acetamidase) is then upregulated up to 100‐fold, such that the enzyme accounts for up to 10% of the total protein content of the bacterial cells.6, 13 As such, the genomic region encoding the amiCADS locus forms the basis of mycobacterial expression vectors under the control of the acetamidase regulon, where the gene amiE has been substituted by a target gene of interest (Fig. 1(B)). This acetamidase regulon has been broadly used as an inducible promoter system on an extra chromosomal plasmid. The recombinant expression vector series, such as pJEM, pSD, or pMyNT were developed, based on the aforementioned vector pMV260.3, 5, 6, 9 However, due to the complexity of the acetamidase regulon, the vector alone is close to 7 kbp in size. The manipulation of the shuttle vector is carried out in E. coli, where the huge size of the vector backbone restricts its usability and stability.14

Figure 1.

Representation of the acetamidase regulon. The genetic compositions of the regulon are schematically represented. The wild type regulon within the M. smegmatis chromosome (A), that within the expression vectors pMyNT (B), that of pMyCA (C), and that of pMyA (D). Bold grey arrows depict the ORFs with the direction of transcription indicated; the gene names are included (black) together with the respective loci numbers of M. smegmatis genes (red). The arrows are connected with a line with the numbers of intergenic base pairs shown in blue. The location of the three promoters within the acetamidase regulon are indicated with small arrows. amiE encodes for the acetamidase enzyme.

The size of the shuttle vector becomes more of an issue, when it comes to expressing a complete operon with multiple ORFs. In prokaryotes, enzymes involved in the same metabolic cascade are often encoded in multi‐cistronic operons. The gene products of such multi‐cistronic operons typically form protein interaction networks and multi‐protein complexes.3, 15 The genome of M. tuberculosis comprises 4.4 Mbp encoding 4000 ORFs, with about half of them being organized within multi‐cistronic operons.2, 15 In earlier experiments, for example, we could show that the co‐expression of EsxA and EsxB from M. tuberculosis was essential to form native and functional complexes3 in contrast to nonsoluble products that were obtained when the proteins were expressed individually.16

The aim of the current work was to construct an expression vector with a minimized acetamidase regulon that allows a controlled induction, and which can be employed with M. smegmatis as an expression host for the study of mycobacterial targets. Here, we present the generation of a new mycobacterial expression vector, designated as pMyCA, which possesses a truncated but fully functional acetamidase regulon (Fig. 1(C)). For this, we performed several gene deletions in the acetamidase regulon found within pMyNT, a derivative of the pSD vector series.9 Among these, the deletion of the amiDS operon resulted in a smaller, functional protein expression vector with induction capacity, which is comparable to previously published data.3 Furthermore, we show the characterization of the vector pMyCA, concerning the growth rate and the protein expression capacity in comparison to the vector pMyNT, which contains the full‐length acetamidase regulon.

Results

Cloning the expression vector pMyCA with a minimized acetamidase regulon by deleting the two ORFs amiS and amiD

The shuttle vector series pSD and pMyNT contain 2704 bp of the acetamidase regulon, comprising the genes amiC‐A‐D‐S (Fig. 1(B)). We carried out several gene deletions within the acetamidase regulon, including a double deletion ΔamiDS and a triple deletion ΔamiCDS (Fig. 1(C,D)). The double gene deletion of ΔamiD and ΔamiS, deleting 1322 bp, was carried out using the specific primer pair ΔamiDS‐BsrGI F/ΔamiDS‐BsrGI R (Table 1) and restriction‐free cloning technology.17, 18 The resulting expression vector pMyCA, containing amiC and amiA (hence the name) was verified by sequencing (Fig. 2). To further evaluate the vector pMyCA, we introduced the mCherry gene into the expression cassette of the vector, resulting in pMyCA‐mCherry. We deposited the plasmids to the Addgene repository (http://www.addgene.org), and they are available under the plasmid IDs # 84268 for pMyCA and #84272 for pMyCA‐mCherry. The vector with the triple deletion ΔamiCDS, termed pMyA, was, however, not functional, resulting in a noninducible expression construct (Supporting Information, Fig. S5).

Table 1.

PCR oligonucleotides for the generation of recombinant pMyCA plasmid

Primer Designation	Nucleotide Sequence
ΔamiDS‐BsrGI F	5′‐TACAAGTGTACA TAAGAGAAAGGGAGTCCACCATGAAGCA‐3′
ΔamiDS‐BsrGI R	5′‐TACAAGTGTACA GGTCACCCCTTTCCATTCACCCGA‐3′
mCherry LIC F	5′‐CGAGAACCTGTACTTCCAGGGC ATGGTGAGCAAGGGCGAG‐3′
mCherry LIC R	5′‐GCCTGGCAGTCGATCGTACG TTACTTGTACAGCTCGTCC‐3′
ΔamiC‐AflII F	5′‐TAGAGACCTTAAG TCCAGATCACCGTCGATCCCGTGT‐3′
ΔamiC‐AflII R	5′‐AGAGACCTTAAG GACCGCGTCACTTCTTTATCTAGATTTA‐3′

Gene specific sequences are shown in bold; vector specific sequences are in italic; BsrGI and AflII recognition sites are underlined.

Figure 2.

Plasmid maps of the pMyNT (A) and pMyCA (B) expression vectors. The plasmid vectors consist of an inducible acetamidase regulon (amiC‐A‐D‐S for pMyNT and amiC‐A for pMyCA), a hygromycin B resistance marker (hygR), an N‐terminal fusion His6‐tag followed by the TEV protease recognition sequence (TEV site) and the respective origins of replication (ColE1 and OriM) for propagation in E. coli and M. smegmatis. A multicloning site (MCS) is present on each vector; the sequence is shown in (C).

Mycobacterium smegmatis transformed with pMyCA have similar growth features as with the vector pMyNT

Growth curves were recorded in triplicate for cultures transformed with the expression vectors pMyNT and pMyCA, both with or without (“empty” vector) a gene of interest. The growth rates of both engineered bacterial strains transformed with the respective empty vectors are comparable (Fig. 3): after an initial lag phase of about 20 h post‐inoculation, both cultures entered the logarithmic phase (log phase). Logarithmic growth was observed until 35 h post‐inoculation. To follow protein expression post‐induction more readily, we used the marker protein mCherry. Due to the interference of the mCherry protein at the wavelength 600 nm (Fig. S4), the growth curves of bacteria transformed with expression vectors including the mCherry gene were monitored at 650 nm. In general, the growth rates of the bacterial cultures were similar plus and minus the mCherry gene. The main difference was that the empty vectors entered the log phase about 4 h earlier than those transformed with the protein of interest, although the end of the log phase remained constant at about 35 h post‐inoculation.

Figure 3.

Growth curves of M. smegmatis transformed with empty vectors pMyNT (black) and pMyCA (orange) measured at OD_600nm, together with the pMyNT‐mCherry (brown) and pMyCA‐mCherry (pink) expression constructs measured at OD_650nm. The time point of induction of protein production by addition of acetamide is indicated. The OD values of the stationary phase differ significantly between pMyCA‐ and pMyNT‐mCherry constructs with 1.9 (CID₉₅: 0.63, 3.24). The results are shown as mean ± standard deviation (n = 3).

We subjected the data of the growth curves to a logistic population growth model19 and to statistical analyses.32 These showed that the optical density (OD) values of the stationary phases of the mCherry cultures using the two different vectors were significantly different: The truncated version (pMyCA‐mCherry) reached the stationary phase with a mean OD_650nm value of 7.7, with a 95% confidence interval (CI₉₅) from 6.6 to 8.7, subsequently stated as 7.7 (CI₉₅: 6.6, 8.7). In comparison, the culture using pMyNT‐mCherry with the full‐length regulon had an average OD_650nm value of 5.8 (CI₉₅: 5.3, 6.2) at the stationary phase. To determine the significance of the difference in stationary growth features of the two vectors, we computed the 95% corresponding CI values for the difference between the two cultures (CID₉₅). The mean difference of OD_650nm was 1.9 with the CID₉₅ interval ranging from 0.63 to 3.24, subsequently referred to as 1.9 (CID₉₅: 0.63, 3.24). This shows that, indeed, they are significantly different, meaning that 10–15 h after induction the yield of cell mass from the bacterial culture transformed with the truncated regulon is around 1.4‐fold higher compared to the culture containing of pMyNT‐mCherry transformants. The statistical treatment of the stationary phases of the control cultures transformed with empty vectors showed a comparable growth behaviour with a mean OD_600nm difference of 0.89 (CID₉₅: −0.63, 2.37).

The minimized acetamidase regulon, consisting of amiC and amiA, is functional for recombinant protein production

The new pMyCA vector has a 1322 bp truncation in the acetamidase regulon, which corresponds to the two genes amiD and amiS, in addition to the acetamidase gene amiE, which is also absent in the pMyNT vector. We used mCherry from Discosoma sp or the esxBA pair from M. tuberculosis (H37Rv) as target proteins to monitor the effect of these truncations on growth and protein expression rates. Lysates of bacteria transformed with the empty vectors or expression constructs were examined using an SDS‐PAGE 24 h post‐induction (Fig. 4). Cell extracts from the same mass of cells were loaded on each lane so that the results are comparable. Induction with acetamide resulted in an overall higher production of proteins. This was also observed for M. smegmatis wild type cells, arising from an induction of the endogenous acetamidase regulon and the bacterial response to the induction (Fig. 4, lanes 2 and 3), confirming published results.11 The band position corresponding to acetamidase is marked with an arrowhead (AmiE, 43 kDa, Figs. 4, S1). The expression vectors containing the recombinant genes clearly showed overexpression for both, mCherry (His₆‐mCherry apparent molecular weight (MW) 35 kDa, Fig. 4, lanes 6 and 7, Fig. S2) and the EsxB:EsxA protein complex (His₆‐EsxB apparent MW 14 kDa and EsxA apparent MW 6 kDa, Fig. 4, lanes 8 and 9, Fig. S3). Furthermore, similar to the plasmid pMyNT,3 pMyCA also supported the bi‐cistronic expression of EsxB and EsxA in an equimolar ratio (Fig. 4, lanes 8 and 9).

Figure 4.

SDS‐PAGE gel showing the level of the recombinant protein production in the pMyNT and pMyCA vectors. Protein marker, lane 1; lysates from M. smegmatis cultures, wild type noninduced control and post‐induction, lanes 2 and 3, respectively; transformed with the empty vectors pMyNT, lane 4; and pMyCA, lane 5; with the expression vectors pMyNT‐mCherry, lane 6; pMyCA‐mCherry, lane 7; pMyNT‐esxBA lane 8; and pMyCA‐esxBA, lane 9. The positions of different bands are indicated on the right side of the gel: the arrowhead points to the bands corresponding to AmiE, the asterisk to the apparent position of the His6‐mCherry‐band and the circles indicate His6‐EsxB (○) and EsxA (•), respectively.

The cells transformed with the vector possessing the truncated acetamidase regulon express about the same amount of recombinant protein per gram cells as those possessing the full‐length regulon

To monitor any differences in recombinant protein production resulting from the ΔamiD‐S deletion, we chose to use the red fluorescent protein mCherry from Discosoma sp. as a reporter of acetamidase regulon activity. The expression of mCherry in pMyNT and pMyCA was monitored as described in the Methods section. In brief, absorption spectra of mCherry from the cleared lysates were measured using a spectrophotometer and quantified using the mCherry absorption maximum20 λ_max = 587 nm with the molar extinction coefficient of ɛ = 72,000 M⁻¹ cm⁻¹. All values were background corrected. We observed a similar level of mCherry expression in the construct with the truncated acetamidase regulon (pMyCA‐mCherry) as with its parent plasmid pMyNT‐mCherry (Fig. 5). However, a notable difference with the pMyNT‐mCherry transformed bacteria was the presence of leaky mCherry expression in the absence of the inducer. In the case of pMyCA‐mCherry, the expression was controlled more tightly. The results are displayed as single data points with their corresponding mean ± SD (n = 6) (Fig. 5).

Figure 5.

Comparison of mCherry protein expression using pMyNT (circles) or pMyCA (squares) vector background. All data points of sextuples are shown grouped in categories of the four different cultures, M. smegmatis transformed either with the vector pMyNT or pMyCA and both either induced (ind, red) or not induced (ni, blue). The semi‐logarithmic representation allows the visualization of the very low leaky expression level of pMyCA‐mCherry without induction (blue squares), which amounts one magnitude less than that of pMyNT‐mCherry (blue circles). The levels of expression obtained with induction are shown in red. The quantities are given as total amount of protein expressed (mg) per gram of M. smegmatis cells (wet weight). The amount of expressed mCherry without induction is significantly (*) lower when using the pMyCA vector with 0.01 mg/g cells (CID₉₅: 0.006, 0.013). In contrast, the expressed amount of mCherry post‐induction per g wet cell pellet is not significantly different (ns) with 0.05 mg/g cells (CID₉₅: −0.056, 0.148). The mean values ± standard deviation of the results are displayed in black (n = 6).

These results can also be observed with the noninduced cultures transformed with vectors including the mCherry gene when samples are examined on an SDS gel (Fig. S2 top, lanes 2–7, Fig. S5). Figure S5 clearly shows the leaky expression of mCherry using the pMyNT vector. In addition, we have subjected the gel to fluorescence detection of mCherry (Fig. S2). The in‐gel fluorescence detection exhibited a clear difference in the fluorescence intensities between the noninduced pMyNT‐mCherry (lanes 2–4) and pMyCA‐mCherry (lanes 5–7) lysates, thereby demonstrating a tighter expression regulation by pMyCA. Under the influence of the complete acetamidase regulon, the induced pMyNT generates 1.14 mg/g cells ± 0.06 mg/g cells (n = 6, CI₉₅: 1.11, 1.16) of mCherry. In comparison, the truncated acetamidase regulon in pMyCA‐mCherry produced 1.09 mg/g cells ± 0.09 mg/g cells (n = 6, CI₉₅: 1.05, 1.13) mCherry. A statistical analysis revealed that indeed the deviation of the expression levels of both vectors are small with 0.05 mg/g cells (CID₉₅: −0.056, 0.148). These results suggest that deletion of the ΔamiD‐S genes has no measurable effect on protein expression per gram cells. However, due to the higher OD values reached in the stationary phase, the yield of the cell mass is higher, therefore, also the protein yield is much higher per liter culture when using the minimized vector pMyCA. In the case of mCherry, the yield is about 1.4 fold using pMyCA versus pMyNT (Fig. 3, Table S3). These findings are further evidenced from the protein expression profiles of mCherry (Fig. 4, lanes 6 and 7, Fig. S2) and the EsxB:EsxA protein complex (Fig. 4, lanes 8 and 9, Fig. S3). The supplement figures show the expression experiments in triplicates. Overexpression of the N‐terminal his‐tagged mCherry (His₆‐mCherry, app. MW 35 kDa) can be clearly observed in all lanes where induction took place, both in the Coomassie stained gel and by fluorescence detection (Fig. S2). A comparison of the mCherry expression levels per gram of cell mass for the pMyCA and pMyNT constructs resulted in similar figures. The same applies for the vectors carrying the genes for the EsxB:EsxA complex (His₆‐EsxB app. MW 14 kDa and EsxA app. MW 6 kDa). pMyNT‐esxBA and pMyCA‐esxBA express their target genes in comparable amounts.

Discussion

Although extrachromosomal plasmids with huge sizes of >40 kb, such as cosmids, can be maintained in prokaryotes with the help of bacteriophages,14 the size of expression vectors for recombinant proteins is typically limited to around 5 kb.21 Shuttle vectors with the inducible acetamidase regulon have been successfully used to produce recombinant proteins using M. smegmatis.3, 6, 9, 22 However, all shuttle vectors developed to date are larger than 7 kb, depending on the vector series. The pJAM series vectors, for example, are 9.4 kb6 whereas pSD derivatives are 7 kb.3, 9 This is due to the presence of two origins of replication and because of the acetamidase regulon having a size of >2.5 kb. The two origins, oriE and oriM, are required for propagation in both, E. coli and M. smegmatis. In this study, we succeeded in minimizing the acetamidase regulon on the plasmid while maintaining its inducibility. The resulting plasmid pMyCA exhibits a truncation of more than 1300 bp and contains only the amiA and amiC genes of the acetamidase regulon. We not only showed that the truncated acetamidase regulon is still functional but that its inducibility is more controllable than that of the parent vector pMyNT. These results agree well with previous findings, which indicated a negative regulatory function for AmiA,12 and those of protein expression in amiA‐amiC knockout strains.11 Even though our newly designed pMyCA plasmid does not include the amiD and amiS genes, they are chromosomally encoded in the expression host M. smegmatis. Therefore, it is well possible that endogenously expressed AmiD and AmiS compensate for their lack in the plasmid upon induction. Furthermore, bacteria transformed with pMyCA grow better than transformants with pMyNT, reaching higher stationary cell densities. These growth features can be observed for both, empty vectors as well as for expression vectors (Fig. 3). It is possible that bacteria grow faster due to the lower number of nucleotides which need to be replicated. However, it is known that the DNA polymerase has the capacity to replicate about >600 bp/s,23 therefore, it cannot completely explain the observed phenomenon. Subsequent biological and functional analyses of the acetamidase regulon will lead to more detailed insights into the observed phenotype.

Apart from its large size, another disadvantage of using the acetamidase regulon has been the basal expression of the target protein in the absence of an inducer.8 To test if the new pMyCA vector showed less basal expression than the pre‐existing pMyNT vectors, we employed the fluorescence protein mCherry. Using mCherry as a marker, we could clearly detect low‐level expression of mCherry in the noninduced pMyNT‐mCherry transformed cultures, both spectroscopically and on an SDS‐PAGE gel. The cultures containing pMyCA‐mCherry showed a tighter control of protein expression with the new truncated acetamidase regulon. It should be noted that the fluorescent bands of mCherry in the lysate samples of noninduced cultures (Fig. S2) could only be visualized using an exposure time of 40 s with a light sensitivity ISO‐800, indicating a very low concentration of mCherry. In‐gel fluorescence detection of mCherry showed some bands with higher apparent MWs up to 43 kDa (Fig. S2). We attribute this observation to some oligomerization due to the nonreducing SDS buffer condition, since the protein samples were not boiled before being submitted to electrophoresis.24

There are two possible explanations for the tighter expression control observed in pMyCA. First, in the process of deleting the amiS gene, a part of the weak constitutive promoter P₃ was deleted. Although the exact sequence of P₃ has not been published, it has been located in the intergenic DNA sequence between amiS and amiE.11 Using the promoter prediction server phiSITE,25 we could identify a possible sequence for P₃. Interestingly, the −35 region of the P₃ promoter overlaps with the coding region of amiS (Supplementary List 1). This would mean that by deleting amiS, the gene of interest in the vector pMyCA is only under the control of the strong P₂ promoter, which is tightly suppressed by AmiA. Hence, the lower basal expression. Second, it might be a dose effect of the expressed suppressor AmiA, which binds to the promoter P₂ downstream of amiA. AmiA is constitutively expressed under the control of the P₁ promoter and binds to the strong P₂ promoter. These are present genomically, as well as on the plasmids (pMyCA and pMyNT). It has been reported that oriM of pAL5000 maintains about five copies of the plasmids per mycobacterial cell.5 This suggests that under the noninducing condition the concentration of AmiA in the cell might not be sufficient to suppress all P₂ sites, thereby showing the observed basal expression (Figs. 5 and S2).

In summary, this contribution provides a new vector tool suitable for producing mycobacterial recombinant proteins. Due to its reduced size compared to other available mycobacterial vectors, pMyCA has the potential to express a number of target proteins that cannot be expressed with the currently available vectors. Together with the previously reported results of heterologous expression using a complete natural operon, this vector now gives researchers the opportunity to study poly‐cistronic operons and express higher‐oligomeric complexes that are commonly found in mycobacteria.3 In addition, M. smegmatis transformed with pMyCA‐mCherry can be employed as an expression control experiment, which can be monitored visually in‐situ without the necessity of any additional detection methods. We routinely perform a parallel control experiment to avoid any unsuccessful expression trials due to the inactive chemical acetamide, which is known to be a hygroscopic and a relatively instable compound, prone to undergo hydrolysis.26

Materials and Methods

Strains and media

Cloning of the plasmids was performed using E. coli XL10 Gold (Agilent Technologies) following the heat‐shock transformation protocol.27 Transformed bacteria were plated onto an LB‐agar (Luria broth) plate (Sigma Life Science, Steinheim, Germany) supplemented with 100 µg/mL Hygromycin B (InvivoGen, Toulouse, France). Plasmid DNA was purified from clones and the positive clones were selected upon DNA sequence verification.

Mycobacterium smegmatis mc²155 was transformed with positive vector DNA by employing the electroporation method,28 and the transformants were cultured on Middlebrook 7H10 agar (Fluka^® Analytical) supplemented with 0.02% Tween^® 80 (Sigma Life Science), 8 g/L bovine serum albumin (BSA Fraction V, Carl Roth), 11.1 mM d‐glucose (Sigma Life Science), 13.9 mM NaCl and 100 µg/mL hygromycin B. Liquid M. smegmatis cultures were grown in 7H9 expression medium (Middlebrook 7H9 medium (Fluka Analytical^®) supplemented with 0.05% (v/v) tween 80, 0.2% (v/v) glycerol, 1.11 mM d‐glucose and 100 µg/mL hygromycin B). The doubling time of M. smegmatis is around 3–4 h.29 Wild type M. smegmatis mc²155 was grown in 7H9 expression medium without antibiotics.

Construction of the pMyCA and pMyA expression vectors

The new mycobacterial expression vector, designated as pMyCA (Fig. 2), was derived from the parent plasmid pMyNT,3 a pSD31 derivate,9 by deleting the amiD‐S operon within the acetamidase regulon. The linearized pMyCA was amplified by polymerase chain reaction (PCR) using the ΔamiD‐S‐BsrGI primer pair (Table 1). The primer pair was designed with a BsrGI restriction recognition site (BsrGI‐HF, New England Biolabs^®) for the circularization of the linear PCR product. The plasmid pMyCA‐esxBA, which contains the coding sequence for the cfp10‐esat6 (esxB/Rv3874 and esxA/Rv3875) operon from M. tuberculosis H37Rv was generated from pMyNT‐esxBA3, 30 using the same ΔamiDS‐BsrGI primer pair.

Because of an internal BsrGI restriction site within the coding sequence of mCherry, mCherry from Discosoma sp. was cloned into pMyNT and pMyCA using a restriction‐free cloning method.18 For this, a chimeric PCR primer pair was designed for the amplification of the mCherry gene (mCherry LIC F/R, Table 1). The PCR product is flanked by overhangs complementary to the vector backbone that allow for a specific insertion into the target plasmid through a secondary PCR.

The pMyA‐mCherry expression vector was derived from pMyCA‐mCherry by deletion of the complete amiC ORF using the ΔamiC‐AflII primer pair (Table 1). The primer pair was designed accordingly with an AflII recognition site for circularization of the linearized PCR product.

All cloning deletion experiments were verified by sequencing using the service of LGC Genomics GmbH employing the primers listed in Table S2. The complete sequences of the new vectors are provided as separated sequence data in fasta format (Supporting Information: pMyCA and pMyCA‐mCherry).

Monitoring the bacterial growth of the transformed M. smegmatis with the expression plasmids with and without target genes: pMyNT and pMyCA

The growth curves were recorded in triplicates. Prior to inoculation of the working cultures, the saturated precultures were prepared; in brief, the precultures of M. smegmatis were made by inoculating 3 mL of Middlebrook 7H9 medium, using a single colony. The working cultures consisted of 200 mL of Middlebrook 7H9 medium inoculated with 1 mL of the saturated precultures. The growth features of M. smegmatis cultures transformed either with the empty vectors, pMyNT or pMyCA, or with the expression constructs, either vector with the mCherry‐insert, were compared. The initial OD_600nm (Ultrospec 100 pro, Amersham Biosciences) was 0.02. The growth curves were recorded over a cultivation period of 48 h. From 18 h post‐inoculation, the culture samples had to be diluted 1:10 for photometric measurements. The growth rates were also monitored under the induction condition; however, the optical density was measured at a wavelength of 650 nm, due to the absorption interference by mCherry at the wavelength 600 nm (Fig. S4). For induction, the cultures were supplemented with 34 mM acetamide at 1 OD_650nm. All cultures were incubated at 37°C and constant shaking at 200 rpm (Shaker Incubator, New Brunswick Scientific).

Expression of recombinant proteins in M. smegmatis and bacterial lysis

One milliliter precultures were employed to inoculate 10 mL working cultures in 7H9 medium. Protein expression was induced at early log‐phase (equivalent to OD₆₀₀ near 1.0) using 34 mM acetamide and an additional incubation period of 24 h post‐induction. The bacteria were harvested by centrifugation at 3500g and 4°C for 10 min and washed once with 1 mL ice‐cold lysis buffer (pH 7.4) containing 20 mM sodium phosphate, 300 mM NaCl, 1 M urea and 20 mM imidazole.

Mycobacterium smegmatis cell pellets were suspended in cold lysis buffer at a ratio of 5 mL buffer per gram wet weight of the bacterial pellet. The buffer additionally included a mix of AEBSF [4(2‐Aminoethyl)benzenesulfonyl fluorid hydrochlorid] 104 µM, leupeptin 10.5 µM and E‐64 1.4 µM (all from Carl Roth) to inhibit bacterial proteases. The bacterial suspensions were then lysed by sonication with a LabSonic 1515 sonicator (B. Braun, Melsungen, Germany) for 1 min at 50 W. Nonsoluble components were separated from the lysate by centrifugation at 20,000g and 4°C for 30 min (Eppendorf Centrifuge 5417R). Afterwards, the clarified supernatants were filtered through a 0.45 µm cellulose acetate membrane syringe filter (Carl Roth). The proteins were visualized by either 15% or 10% SDS‐PAGE gels, run at 120 V for 70 min and stained with colloidal coomassie stain.31 Protein samples for electrophoresis containing mCherry were not heat‐denatured to prohibit the temperature dependent hydrolysis of the main chain acylimine linkage.32

Fluorescence detection of mCherry in SDS‐PAGE gel

The protein content of the lysate samples of noninduced and induced pMyCA‐mCherry and pMyNT‐mCherry cultures (each n = 3) were electrophoretically separated on a 15% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS‐PAGE) gel. Prior to the Coomassie staining, the fluorescence of mCherry was detected using an Olympus E‐410 digital camera with an Olympus Digital objective (17.5–45 mm, 1:3.5–5.8). Within the light‐tight chamber, the gel was illuminated with green light emitting diode light (λ = 530 nm ± 30 nm) and the emission was filtered with an ET685/70 bandpass filter (Chroma Technology Corporation, Vermont) mounted in front of the camera objective. The image of the SDS‐PAGE gel with the samples of induced cultures was taken with a sensitivity of ISO‐400, F/5.1 and with an exposure time of 8 s. The settings for the gel with the noninduced samples were ISO‐800, F/5 and with an exposure time of 40 s.

Quantification of promotor activity

The acetamidase promotor activity from pMyNT and pMyCA was assessed by the comparison of their respective expression of mCherry. Optical densities from the high spin mCherry lysates were measured using a NanoDrop 1000 (Peqlab Biotechnologie GmbH). Background absorption was subtracted to obtain the absorption from mCherry only. The total mCherry quantities (given as milligram mCherry per gram bacteria cell) for each sample were calculated from the background corrected optical density at the wavelength 587 nm, where mCherry has its absorption maximum with ɛ = 72,000 M⁻¹ cm⁻¹. All experiments were carried out in sextuplicates and the results are given as mean ± standard deviation with the CI₉₅ of the mean in brackets.

Statistical analysis

The growth curves of M. smegmatis were fitted using the nonlinear equation $y=A·\mathrm{e}\mathrm{x}\mathrm{p}\left\{\left(-\mathrm{e}\mathrm{x}\mathrm{p}\left[({\mu }_m·e/A)·\left(\lambda -t\right)+1\right]\right\} \right)$with the Marquardt algorithm,19 where $A$ denotes the maximal OD value at the stationary phase, ${\mu }_m$ the slope of the logarithmic growth phase, and $\lambda$ the lag time of each curve, which is determined as the x‐axis intercept of the logarithmic growth tangent.19 We subjected the obtained data from the growth curves and expression studies to confidence interval comparison at the CI₉₅ using the equation: ${\mathrm{C}\mathrm{I}}_{95}=M\pm S{E}_{\mathrm{M}}·t_{n-1;0.975}$, where $M$ is the mean value, $S{E}_{\mathrm{M}}$ the standard error of mean, and $t_{n-1;0.975}$ is the $t$‐distribution coefficient (32). The values are given in the form of “average value (CI₉₅ lower limit of the interval, upper limit thereof).” To test for statistical significance at the 95% confidence level, we computed the confidence interval for the difference (CID₉₅) between two sample means. The following equation employed:

${\mathrm{C}\mathrm{I}\mathrm{D}}_{95}=M_{1/2}\pm S{E}_{\mathrm{D}}·t_{n_1+n_2-2;0.975}$, where $M_{1/2}$ is the mean values of two independent data sets, and $S{E}_{\mathrm{D}}$ the estimated standard deviation thereof.33 If a CID₉₅ includes the Null‐value, it is considered as not statistically significant or meaningful.

CONFLICT OF INTEREST

The authors declare that they have no conflict of interest.

Supporting information

Supporting Information

Supporting Information

Supporting Information

Supporting Information