Construction of shuttle vectors for genetic manipulation and molecular analysis of Mycobacteria

Abstract

Two novel shuttle vectors for mycobacteria are described which have been derived from the expression system pSD5 developed in our laboratory. Plasmid pSD5B is a promoter-selection vector containing a promoterless lacZ gene and allows the identification of mycobacterial promoters by the blue colour of the colonies on solid media containing XGal. Moreover, the chronological order of appearance of blue colonies and intensity of colour provide a qualitative index of transcriptional strengths of the cloned promoters. Plasmid pSD5C has been designed to construct mycobacterial genomic libraries and express the cloned DNA inserts as fusion proteins with maltose binding protein in mycobacteria. Libraries in pSD5C provide feasibility for their screening with either DNA probes or specific antisera for identifying the genes of interest and for isolation of specific genetic loci by complementation of Escherichia coli and mycobacterial mutants. These vectors combine the ease of working in E. coli with the advantage of directly propagating them in mycobacteria without further manipulations. Finally, we demonstrate that these vectors function efficiently both in fast growing Mycobacterium smegmatis and slow growing mycobacteria including Mycobacterium tuberculosis and Mycobacterium bovis BCG.

1. Introduction

Tuberculosis (TB), the cause of the largest number of deaths from a single infectious agent worldwide, continues to present serious health concerns around the globe. More than 50 000 people die of tuberculosis every week and one third of the world’s population is estimated to be infected with the tubercle bacilli asymptomatically (Kochi, 1994). While 95% of all cases of tuberculosis occur in developing countries, the developed world too has seen its revival due to the spread of the human immunodeficiency virus. Infection with AIDS virus is associated with elevated susceptibility to tuberculosis and increased mortality and morbidity (Chaisson and Slutkin, 1989). The emergence of multi-drug resistant strains of Mycobacterium tuberculosis has further aggravated the problem (Frieden et al., 1993).

The present challenge for mycobacteriologists is to apply molecular genetic approaches to understand the biology of mycobacteria and utilize this knowledge for the development of more effective prophylactic, diagnostic and curative methods to combat mycobacterial infections. The study of molecular biology of mycobacteria in the past was rendered difficult due to reasons such as slow growth rate and high GC content of their genomes and the fact that they remained refractory to transformation with DNA due to their tough cell wall and lack of suitable plasmid vectors for gene transfer. The last decade, however, has witnessed considerable progress in this field and genetic tools for cloning and expression of mycobacterial genes have become available. Bacteriophage λgt11 was the first to be used as a vehicle for the construction of mycobacterial genomic libraries that permitted the use of specific antisera to isolate immunoreactive antigens (Young et al., 1985, 1987). Subsequently these libraries were used to perform complementation of Escherichia coli mutants to identify mycobacterial genes (Garbe et al., 1990; Aldovini et al., 1993; Anderson and Hansen, 1993; Cirillo et al., 1994). Isolation of high-efficiency-transformation strains of the fast growing M. smegmatis (Snapper et al., 1990) and development of electroporation to transform mycobacteria (Snapper et al., 1988) were crucial achievements in our ability to express genes in mycobacteria with a variety of shuttle vectors that were based on the plasmid pAL5000 from M. fortuitum (Labidi et al., 1985; Snapper et al., 1988; Ranes et al., 1990; Stover et al., 1991). Plasmids and integration proficient vectors based on sequences derived from mycobacteriophages have enabled stable introduction and expression of genes in mycobacteria (Jacobs et al., 1987; Lee et al., 1991; Stover et al., 1991).

We had designed an expression system pSD5 for mycobacteria which provides varied levels of expression of a gene by changing the mycobacterial promoter regulating its expression (S.K.D., S.J., D.K. and A.K.T., unpublished). We report here the construction of two vectors, pSD5B and pSD5C, which have been derived from pSD5 and can be used for the isolation of transcriptionally active DNA fragments from mycobacteria and the construction of genomic libraries of mycobacteria respectively. We describe the salient features and applications of these vectors that overcome some of the limitations of the mycobacterial vectors described previously.

2. Experimental and discussion

2.1. Isolation of mycobacterial promoters based on blue-white selection

The vector pSD5B was constructed to isolate promoter sequences from the mycobacterial genome (Fig. 1). Promoter cloning sites XbaI and SphI are preceded by the transcriptional terminator TER3 to prevent any readthrough transcription from the mycobacterial origin of replication (oriM). Translational terminators in all three reading frames between the promoter cloning sites and IacZ are present to uncouple transcriptional initiation from translational initiation. A transcriptional terminator, TER4, derived from fd phage is present downstream from IacZ to facilitate cloning of strong promoters (Gentz et al., 1981).

Fig. 1.

Strategy for construction of promoter selection vector pSD5B. The vector pMC1871 (Pharmacia Fine Chemicals, Uppsala, Sweden) was digested with SmaI + ScaI yielding 3.3-kb and 4.2-kb fragments. The 3.3-kb fragment containing the promoterless 1acZ gene was purified and cloned into ScaI-digested, dephosphorylated pSD5 in the orientation as shown. P15A ori and oriM are the origins of DNA replication for propagation in E. coli and mycobacteria, respectively. Promoter cloning sites, XbaI and SphI, transcriptional terminator TER3 to prevent read-through transcription from oriM, translational terminators in all three reading frames between the promoter cloning sites and lacZ, and the transcriptional terminator TER4 to facilitate cloning of strong promoters are all shown in the figure.

A partial promoter library using mycobacterial DNA was prepared in pSD5B (see legend to Fig. 2). The clones exhibiting promoter activities were identified by the appearance of blue coloured mycobacterial colonies on LB agar containing XGal, due to expression of βGal in these bacteria. Using pSD5B, we were able to isolate mycobacterial promoters with a frequency of 10%. The plasmid DNA purified from promoter clones was subjected to restriction digestion with SphI and it was observed that all the plasmids carried DNA inserts ranging from 0.1 to 0.5 kb in size (Fig. 2). The clones were assigned serial numbers in the order of their acquisition of blue colour. βGal activities in the cell-free extracts prepared from individual clones indicated that the promoter clones 1–17 exhibited activities ranging between 2879 and 135 nmol/min/mg of protein, thus providing a 20-fold variation in promoter strengths (Fig. 3A). The extracts were also analysed by 0.1% SDS-7.5% PAGE and probed with anti-βGal Ab (Fig. 3B). The expression profile correlated with βGal activities from respective clones thereby confirming the range of promoter strengths. To assess the functioning of pSD5B in slow-growing species of mycobacteria, six promoter clones (1, 2, 5, 8, 10 and 14 in Fig. 3) were electroporated into M. tuberculosis H₃₇Ra and M. bovis BCG. Transformants of all these six promoters were blue in both species. Cell-free extracts from the promoter clones exhibited βGal activities comparable with those observed in M. smegmatis (data not shown) thus substantiating that the vector pSD5B functions efficiently in slow-growing mycobacteria as well.

Fig. 2.

Restriction analysis of mycobacterial promoters isolated using pSD5B. For isolation of promoters in pSD5B, genomic DNA from M. smegmatis LR222 was digested with SphI, and the fragments from 100 bp to 1 kb were purified. In a typical experiment, 60 ng of purified genomic DNA fragments were ligated to 100 ng of SphI-cleaved, dephosphorylated pSD5B, and electroporated into M. smegmatis LR222. Transformants were selected on LB agar containing 25 µg Km/ml and 40 µg XGal/ml (Life Technologies, Gaithersberg, MD, USA) and assigned numbers in the serial order of appearance of blue colour. Individual transformants were grown in LB medium supplemented with 0.5% glycerol/0.2% Tween-80/25 µg Km/ml to an A_600nm of 2.0. Plasmid DNA was isolated by a modified alkaline SDS method, subjected to restriction digestion with SphI and analysed on a 6% polyacrylamide gel. Lanes: 1, pSD5B digested with SphI; 2–9, plasmid DNA from eight of the promoter clones digested with SphI; M1, phage λ DNA digested with HindIII M2, pGEM-5Z f(+) DNA digested with HaeIII. Sizes of marker bands in kb are indicated on the left for M1, and in bp on the right for M2.

Fig. 3.

Expression of lacZ in the promoter clones isolated using pSD5B. (A) Bar diagram depicting βGal activities of pSD5B promoter clones in M. smegmatis. (B) Western blot analysis with anti-βGal Ab (Sigma, St. Louis, MO, USA). The position of βGal is indicated by an arrow. Methods: Individual promoter clones were grown in LB medium supplemented with 0.5% glycerol/0.2% Tween-80/25 µg Km/ml to an A_600nm of 2.0. Cells were harvested, resuspended in one sixth of the culture volume of 0.25 M Tris · HCl (pH 7.6), disrupted by sonication and centrifuged at 12 000 × g for 10 min at 4°C. Total protein and βGal activities were estimated in cell-free extracts as described by Miller (1972) and expressed as nmol ONPG converted to o-nitrophenol per min per mg of protein. Protein samples for SDS-PAGE were prepared by addition of 4 × loading buffer to a final concentration of 1 × (50 mM Tris · HCl pH 6.8/l% SDS/1% β-mercaptoethanol/10% glycerol/0.1% bromophenol blue), boiled for 10 min, centrifuged at 12 000 × g for 10 min at room temperature, subjected to 0.1% SDS-7.5% PAGE and transferred to nitrocellulose membrane (Schleicher&Schuell, Keene, NH, USA) for probing with Ab.

Selection of promoters based on lacZ (βGal) expression by pSD5B offers an advantage in that cloning of a wide range of promoters is possible without incorporating any bias towards the promoters of a certain range as can occur in vectors using antibiotic resistance-encoding genes as a basis for promoter selection (DasGupta et al., 1993). Besides, the mycobacterial promoters isolated in pSD5B can be directly assessed for their transcriptional strength in E. coli. The presence of ori both for E. coli (P15A) and mycobacteria (oriM) and the gene for kanamycin resistance which functions in E. coli and mycobacteria provide the basis for propagation of this vector for isolation and analysis of promoters in these bacterial species. It has been observed that vectors maintaining a high copy number in E. coli often express the reporter gene in the absence of any promoter sequences due to a large accumulative read-through transcription thus rendering them unsuitable for use in E. coli (Timm et al., 1994). We have circumvented this problem in pSD5B by using the P15A ori which maintains a plasmid copy number of 15–20 in E. coli (Chang and Cohen, 1978) and does not produce any βGal in the absence of promoter sequences (data not shown).

2.2. Construction of pSD5C

In designing the vector pSD5C for expression of genes in mycobacteria, we have taken advantage of the fact that the majority of E. coli promoters function in mycobacteria albeit with efficiencies different from those in E. coli (Bashyam et al., 1996). In pSD5C, gene expression is under the transcriptional control of the Ptac promoter. Thus, expression of cloned DNA fragments in an E. coli strain overexpressing lacI (encoding lac repressor) commences only upon the addition of IPTG which induces expression from Ptac promoter. This promoter, although not induced by addition of IPTG in mycobacteria, can function constitutively at a level much lower than that observed after induction in E. coli. As such, while the induced level of expression in E. coli is high enough to be seen on Coomassie blue stained polyacrylamide gels, that in mycobacteria can be detected using Ab. pSD5C contains the malE gene (encoding maltose-binding protein, MBP) fused to the lacZα fragment downstream from Ptac promoter (Fig. 4). DNA fragments can be cloned in the unique PstI site within lacZα. The transcriptional terminator from fd phage (TER4) present downstream from lacZα prevents any transcription originating from Ptac from interfering with plasmid functions. In addition, the presence of oriM and Km^R gene facilitates selection and expression from pSD5C or its recombinants in mycobacteria.

Fig. 4.

Strategy for construction of the vector pSD5C. pSD5 DNA was digested with XbaI + MluI. The 6.2-kb DNA fragment containing oriM P15A ori and the Km^R gene was purified, dephosphorylated, end repaired and ligated to a 2.1-kb SspI fragment, containing the malE gene fused to IacZα obtained from pMAL-C2 (New England Biolabs, Beverly, MA, USA). The clones were selected with the orientation of malE-IacZα fragment as shown. For other designations, see legend to Fig. 1.

2.3. Application of pSD5C in expression of genes in E. coli and mycobacteria

In E. coli transformed with pSD5C, addition of IPTG induces the expression of the malE-IacZα genes, producing a 55-kDa fusion protein (Fig. 5A and B, lane 2). To examine that all necessary features of pSD5C function appropriately, a fragment encoding CAT (gene cat, containing its own ribosome-binding site and start codon without the promoter region) derived from pSD7 (DasGupta et al., 1993) was cloned in the PstI site of pSD5C to obtain pSD5C-CAT and the expression of cat was analysed in E. coli and M. smegmatis by 0.1% SDS-10% PAGE and Western blotting. Since cat was cloned in a reading frame different from malE, it was not expressed as an MBP fusion protein. Instead, two proteins, corresponding to MBP (38 kDa) and CAT (26 kDa), were detected in induced cultures of E. coli (Fig. 5A, lane 4). Their authenticity was confirmed with anti-MBP Ab (Fig. 5B, lane 4) and anti-CAT Ab (Fig. 5C, lane 4).

Fig. 5.

Expression in E. coli and M. smegmatis using pSD5C. (A) Coomassie blue stained SDS-polyacrylamide gel. (B) Western blot analysis with anti-MBP Ab (New England Biolabs, Beverly, MA, USA). (C) Western blot analysis with anti-CAT Ab. Lanes: A1, B1 and C1, 20 pg, 2 µg and 2 µg protein, respectively, from cell-free extract of E. coli XL1 Blue[pSD5C] uninduced; A2, B2 and C2, 20 µg, 2 µg and 2 µg protein, respectively, from cell-free extract of E. coli XL1 Blue[pSD5C] induced with IPTG; A3, B3 and C3, 20 µg, 2 µg and 2 µg protein, respectively, from cell-free extract of E. coli XL1 Blue[pSD5C-CAT] uninduced; A4, B4 and C4, 20 µg, 2 µg and 2 µg protein, respectively, from cell-free extract of E. coli XL1 Blue[pSD5C-CAT] induced with IPTG; A5, B5 and C5, 20 µg, 100 µg and 100 µg protein, respectively, from cell-free extract of M. smegmatis LR222[pSD5C]; A6, B6 and C6, 20 µg, 100 µg and 100 µg protein, respectively, from cell-free extract of M. smegmatis LR222[pSD5C-CAT]. Molecular masses of marker proteins in A and expressed proteins in B and C are indicated in the left margin. Methods: E. coli XL1 Blue (Stratagene, La Jolla, CA, USA) was transformed with pSD5C-CAT and grown to an A_{600 nm} of 0.6 in LB medium supplemented with 25 µg Km/ml and 10 µg Tc/ml. The culture was then divided into two equal portions. After addition of IPTG (to a final concentration of 50 µM) to one of the portions, both cultures were further incubated for 1 h with shaking at 37°C before harvesting. E. coli XL1 Blue[pSD5C] was subjected to identical treatments and served as a control. For expression in mycobacteria, M. smegmatis was transformed with pSD5C and pSD5C-CAT separately and grown as described for Fig. 3. Cell-free extract from each transformant was prepared as described in the legend to Fig. 3, subjected to 0.1% SDS-10% PAGE and stained with Coomassie brilliant blue R-250 (0.25% in 50% methanol and 10% acetic acid) or transferred to nitrocellulose membrane (Schleicher&Schuell, Keene, NH, USA) for probing with Ab.

Cell-free extracts prepared from M. smegmatis transformed with pSD5C or pSD5C-CAT were also analysed similarly. While Coomassie blue stained SDS-polyacryl-amide gels did not reveal any band due to MBP-LacZα, MBP or CAT (Fig. 5A, lanes 5 and 6), their expression was detected when Ab to MBP (Fig. 5B, lanes 5 and 6) and to CAT (Fig. 5C, lane 6) were used. M. smegmatis transformed with pSD5C-CAT formed colonies on LB agar plates containing 25 µg Km/ml and 20 µg Cm/ml thus confirming that enzymatically active CAT could be expressed in M. smegmatis using pSD5C. Based on the amounts of protein loaded for a comparable expression of MBP-LacZα fusion protein and CAT in E. coli and mycobacteria as detected using Ab (Fig. 5B and C) the constitutive level of expression from Ptac promoter in mycobacteria is approx. l–2% of the expression in E. coli after induction with IPTG.

2.4. Construction of mycobacterial genomic library in pSD5C

A partial genomic library of M. tuberculosis H₃₇Ra was constructed in pSD5C and expression from some of the recombinants tested in E. coli and M. smegmatis. Upon induction with IPTG (Fig. 6A), MBP fusion proteins of various sizes were expressed in eight of the E. coli recombinants. The same recombinant proteins could be detected with anti-MBP Ab when expressed in M. smegmatis (Fig. 6B). The ability of this expression system to express genes both in mycobacteria and E. coli should provide ease of screening the libraries constructed in pSD5C either using nucleic acid probe or a specific Ab in E. coli. The identified clones can be directly electroporated into mycobacteria for expressing the proteins without further manipulations, an advantage not reported for any of the existing plasmid vectors for constructing mycobacterial genomic libraries. Such libraries may be extremely useful in genetic complementation of both E. coli and mycobacterial mutants for isolation of mycobacterial genes that can be identified based on phenotypic selection.

Fig. 6.

Expression of MBP fusion proteins in pSD5C recombinant clones containing mycobacterial genomic DNA fragments. (A) Analysis of MBP fusion proteins on SDS-polyacrylamide gel stained with Coomassie Brilliant Blue. 20 µg total protein from cell-free extracts was loaded in each lane. Lanes: 1, E. coli XL1 Blue[pSD5C] uninduced; 2, E. coli XL1 Blue[pSD5C] induced with IPIG; 3–10, E. coli XL1 Blue[pSD5C] recombinants containing M. tuberculosis H₃₇Ra genomic DNA fragments induced with IPTG. (B) Western blot analysis of MBP fusion proteins with anti-MBP Ab. 100 µg of total protein from cell-free extracts was loaded in each lane. Lanes: 1, M. smegmatis LR222[pSD5B] as control; 2, M. smegmatis[pSD5C]; 3–10, M. smegmatis [pSD5C] recombinants containing M. tuberculosis H₃₇Ra genomic fragments (loaded in the same order as in A). Molecular mass markers (in kDa) are indicated on the left of each panel. Methods: Genomic DNA isolated from M. tuberculosis H₃₇Ra was digested with PstI and fragments ranging between 2 and 7 kb were purified. These were ligated to PstI-digested, dephosphorylated pSD5C, and the ligation mixture was used to transform E. coli XL1 Blue. Transformants were selected on LB agar containing 25 µg Km/ml and 10 µg Tc/ml. For expression studies, individual transformants were grown in LB medium containing Km and Tc to A_600nm of 0.6, and induced with 50 µM IPTG for 1 h with shaking at 37°C. M. smegmatis transformants were grown and their cell-free extracts were prepared and analysed by 0.1% SDS-IO% PAGE, as described in the legend to Fig. 3.

2.5. Use of pSD5C for protein expression in M. tuberculosis and M. bovis BCG

Expression of maIE-lacZα and cat in pSD5C was examined in the slow-growing mycobacterial species M. tuberculosis H₃₇Ra and M. bovis BCG. The 55-kDa MBP-LacZα fusion protein was detected in cell-free extracts of M. smegmatis, M. tuberculosis H₃₇Ra and M. bovis BCG transformants harbouring pSD5C (Fig. 7A, lanes 1, 3 and 5). pSD5C-CAT transformants of M. smegmatis, M. tuberculosis H₃₇Ra and M. bovis BCG, respectively, displayed a 38-kDa protein band due to MBP alone (as described above in Section 2.3) when probed with anti-MBP Ab (Fig. 7A, lanes 2, 4 and 6) and a 26-kDa band of CAT protein when probed with anti-CAT Ab (Fig. 7B, lanes 2, 4 and 6). The ability to express genes in the slow-growing mycobacteria using pSD5C is of extreme importance. Complementation of non-pathogenic mycobacterial species such as M. tuberculosis H₃₇Ra and M. bovis BCG can be carried out with the genomic libraries of their pathogenic counterparts such as M. tuberculosis H₃₇Rv and M. bovis for identifying the genetic determinants responsible for the disease causing ability of the latter. Such a strategy has been employed using the cosmid library of M. tuberculosis H₃₇Rv to complement M. tuberculosis H₃₇Ra and has led to the identification of DNA fragments that provide a growth advantage to the recombinant M. tuberculosis H₃₇Ra in mice (Pascopella et al., 1994).

Fig. 7.

Expression of cat using pSD5C in different mycobacterial species. (A) Western blot analysis with anti-MBP Ab. (B) Western blot analysis with anti-CAT Ab. Lanes: A1 and B1, M. smegmatis[pSD5C]; A2 and B2, M. smegmatis[pSD5C-CAT]; A3 and B3, M. tuberculosis H₃₇Ra[pSD5C]; A4 and B4, M. tuberculosis H₃₇Ra[pSD5C-CAT]; A5 and B5, M. bovis BCG[pSD5C]; A6 and B6, M. bovis BCG[pSD5C-CAT]. Methods: M. tuberculosis H₃₇Ra and M. bovis BCG were separately transformed with pSD5C and pSD5C-CAT as described previously (Bashyam et al., 1996) and transformants were selected on Middlebrook 7H10 agar supplemented with OADC (Difco Laboratories, Detroit, MI, USA), 25 µg Km/ml and 20 µg Cm/ml. Single colonies were grown in Middlebrook 7H9 medium containing ADC (Difco Laboratories, Detroit, MI, USA) and Km to A_600nm of 2.0 and cell-free extracts were prepared and analysed by 0.1% SDS-10% PAGE as described in the legend to Fig. 3. 100 µg total protein from cell-free extracts was loaded in each lane.

Post-translational modifications such as glycosylation and acylation and secretion of proteins that occur specifically in mycobacteria (Garbe et al., 1993; Burlein et al., 1994; Matsuo et al., 1990) cannot be achieved if these genes are expressed in E. coli. pSD5C provides the possibility of expressing such genes in mycobacteria and easy purification of appropriately modified proteins directly from mycobacteria on affinity columns using amylose resins. Due to the presence of factor Xa cleavage site between MBP and the expressed protein, recombinant protein can be cleaved with factor Xa to release pure mycobacterial protein synthesized in its native environment.

3. Conclusions

1. Two shuttle vector systems pSD5B and pSD5C have been constructed for mycobacteria by modifying the expression system pSD5 and are suitable for analysis of gene expression in fast and slow-growing mycobacteria.

2. pSD5B is a promoter selection vector carrying the promoterless lacZ gene that can be used to identify transcriptionally active DNA fragments in mycobacteria based on the blue-white selection.

3. The vector pSD5C is designed for construction of genomic libraries of mycobacteria wherein gene fragments are expressed as MBP fusion roteins under the transcriptional control of the E. coli Ptacpromoter. Protein expression in E. coli can be induced with IPTG, hence, it is possible to screen such genomic libraries in E. coli and the selected clones can be expressed constitutively in mycobacteria without any further molecular manipulations.

4. The presence of MBP as a fusion protein tag in pSD5C would aid in affinity purification of fusion proteins directly from mycobacteria which can be subsequently cleaved with factor Xa to release mycobacterial proteins in their native conformation and with proper posttranslational modifications.

Abbreviations

A

absorbance

Ab

antibody (ies)

βGal

β-galactosidase

CAT

chloramphenicol acetyl transferase

Cm

chloramphenicol

IPTG

isopropyl-β-d-thiogalactopyranoside

Km

kanamycin

LB

Luria-Bertani (medium)

M.

Mycobncterium

MBP

maltose-binding protein

MBP-LacZα

fusion protein of maltose-binding protein with α-fragment of β-galactosidase

ONPG

o-nitrophenyl-α-d-galactopyranoside

ori

origin of DNA replication

PAGE

poly-acrylamide gel electrophoresis

R

resistance

SDS

sodium dodecyl sulfate

Tc

tetracycline

XGal

5-bromo-4-chloro-3-indolyl-β-d-galacto-pyranoside

[]

denotes plasmid-carrier state