Production of recombinant proteins in Mycobacterium smegmatis for structural and functional studies

Abstract

Protein production using recombinant DNA technology has a fundamental impact on our understanding of biology through providing proteins for structural and functional studies. Escherichia coli (E. coli) has been traditionally used as the default expression host to over-express and purify proteins from many different organisms. E. coli does, however, have known shortcomings for obtaining soluble, properly folded proteins suitable for downstream studies. These shortcomings are even more pronounced for the mycobacterial pathogen Mycobacterium tuberculosis, the bacterium that causes tuberculosis, with typically only one third of proteins expressed in E. coli produced as soluble proteins. Mycobacterium smegmatis (M. smegmatis) is a closely related and non-pathogenic species that has been successfully used as an expression host for production of proteins from various mycobacterial species. In this review, we describe the early attempts to produce mycobacterial proteins in alternative expression hosts and then focus on available expression systems in M. smegmatis. The advantages of using M. smegmatis as an expression host, its application in structural biology and some practical aspects of protein production are also discussed. M. smegmatis provides an effective expression platform for enhanced understanding of mycobacterial biology and pathogenesis and for developing novel and better therapeutics and diagnostics.

Introduction

Use of recombinant DNA technologies to produce large quantities of proteins for structural and functional studies is today a routine practice. Considering the substantial role of recombinant protein production in almost all fields of life sciences, the choice of an appropriate expression method is a crucial variable in producing proteins in high quantity and quality.¹ In this regard, both improving existing expression strategies and developing new expression systems are considered as the greatest challenge for structural genomics projects.²

Escherichia coli (E. coli) is the most prevalent microbial factory in biology.³ The major bottleneck, however, in the production of recombinant proteins in E. coli is the formation of insoluble inclusion bodies.³ This has been particularly discouraging in the expression of proteins from mycobacterial species including Mycobacterium tuberculosis (Mtb), the bacterium that causes tuberculosis. For this species, success in soluble expression, even with the best of E. coli strategies, has rarely exceeded 30%.⁴ This is partly attributed to the different G+C content between Mtb (65.6%)⁵ and E. coli (50.8%)⁶ genomes, requiring different machinery for efficient transcription and translation of mycobacterial genes.² Several other factors that may play a role in the inability of E. coli cells to produce soluble and properly folded recombinant proteins, and strategies to overcome those limitations, are discussed elsewhere.³

In recognition of the limitations of E. coli host cells as a “one size fits all” strategy, various alternative expression systems for producing recombinant proteins have been developed.⁷ The central idea is that problems in protein expression can be most effectively addressed by using host organisms more closely related to the organism from which the proteins are derived. There are, for instance, ∼70 different expression hosts reported for protein production in the Protein Data Bank (PDB).⁷ We presume the number of alternative expression hosts could be even larger since not all proteins are produced for structural studies.

Expression Hosts for Mycobacterial Proteins

Publication of the Mtb H37Rv genome sequence in 1998⁵ ignited a widespread interest in the biology and pathogenesis of mycobacteria, enabling functional and structural investigation of proteins from various mycobacterial species (Fig. 1). This led in turn to formation of the Mtb Structural Genomics Consortium (http://www.webtb.org) to “develop a foundation for tuberculosis diagnosis and treatment by determining the three-dimensional structures of proteins”.⁸ However, the success rate of obtaining soluble Mtb proteins from E. coli host cells is far from ideal, with only one third of the proteins being produced in soluble and properly folded form, suitable for structural and functional studies.⁴ To address this problem, different expression hosts have been used for heterologous expression of proteins from Mtb and other pathogenic mycobacterial species.

Figure 1.

Protein structures present in the Protein Data Bank from different mycobacterial species. The numbers indicate discrete entries with <90% sequence identity. For clarity, the figure does not include species for which only two (M. xenopi and M. fortuitum) or one (M. gastri, M. goodie, M. vanbaalenii, M. sp JLS, and M. sp JC1) entry is found.

Baculovirus/insect host cells

The baculovirus/insect cell expression system is one of the most widely used systems for routine production of recombinant proteins.⁹ This expression system was one of the first to be used in attempts to produce Mtb proteins in an alternative expression host¹⁰ and has been used for production of a limited number of mycobacterial proteins.^11,12 This expression system has, however, been overtaken by other expression hosts with closer evolutionary links to mycobacteria.

Streptomyces host cells

Mycobacteria and Streptomyces are members of the Actinobacteria family, and therefore share close evolutionary links. Initial observations that Streptomyces lividans (S. lividans) could utilize the Mycobacterium bovis (M. bovis) BCG translational signals suggested its use as an expression host for mycobacterial genes.¹³ S. lividans is also an efficient micro-organism for expression and secretion of recombinant proteins¹⁴ and has thus been successfully used to express and purify Mtb proteins from the culture medium.¹⁵ S. lividans also supports glycosylation of Mtb proteins,¹⁶ reinforcing its usefulness for heterologous production of proteins from pathogenic mycobacteria.

Mycobacterial host cells

Proteins from pathogenic mycobacterial species have been previously expressed in non-pathogenic mycobacteria, including Mycobacterium smegmatis (M. smegmatis)¹⁷ and Mycobacterium vaccae (M. vaccae).¹⁸ The former, however, is now the species of choice, given its successful application as an efficient expression host (Table 1). In this review, we discuss the various expression systems and strategies developed for protein expression and purification from M. smegmatis host cells.

Table 1.

Examples of Mtb Proteins Expressed in M. smegmatis Expression Host

				Solubility
ORF	Function/annotation	Molecular weight (kDa)	Theoretical pI	Soluble	Insoluble	Oligomeric state	References
Rv0065	VapC1	14.3	4.8	✓			79
Rv0132c	fHMAD	38.4	6.4	✓		Dimer	63, 80
Rv0407	FGD1	36.9	5.1	✓		Dimer	72
Rv0413	MutT3	23.4	4.8		✓		4
Rv0475	HBHA	21.5	9.8	✓			59
Rv0542c	MenE	36.5	5.6		✓		n/a
Rv0617	VapC29	13.9	5.7	✓			79
Rv0901	CP	18.9	10.1				81
Rv0993	GalU	32.3	4.6	✓			82
Rv1096	Possible glycosyl transferase	31	7	✓			83
Rv1160	MutT2	15.1	6.1	✓			4
Rv1186c	PruC	57.5	5.9	✓a			84
Rv1187	PruA	58.8	5.7	✓		Dimer	73
Rv1188	PruB	36.2	8.3	✓a			73
Rv1212c	GlgA	41.5	6.3	✓			82
Rv1213	GlgC	43.8	4.9	✓			82
Rv1601	HisB	22.7	6.8	✓		Monomer	85
Rv1665	PKS11	37.6	5.6	✓		Dimer	71
Rv1886c	FbpB	34.5	5.7	✓		Monomer	86
Rv1980c	MPT64	24.8	4.6	✓			41
Rv1991c	MazF6	12.2	9.9	✓			44
Rv2141c	DapE2	48.1	4.7	✓			4
Rv2220	GlnA1	53.5	4.8	✓		Dodecamer	87–88
Rv2245	KasA	43.2	4.9	✓		Dimer	89
Rv2420c	RsfS	13.8	4.5	✓		Dimer	PDB code 4WCW
Rv2611c	Possible acetyltransferase	35	9.1	✓			4
Rv2903c	LepB	31.8	5.7		✓		n/a
Rv2924c	MutM	31.9	10.3	✓			4
Rv3127	CP	38.4	7.6		✓		4
Rv3177	Possible peroxidase	30.8	8.9		✓		4
Rv3248c	SahH	54.3	4.8	✓			90
Rv3490	OtsA	55.8	6.7	✓			82
Rv3547	Ddn	17.3	9.9	✓a			4
Rv3565	AspB	41	4.6	✓			91
Rv3713	CobQ2	24.1	6.1	✓			4
Rv3772	HisC2	38	6.5	✓			92
Rv3855	EthR	23.7	4.9	✓		Dimer	93
Rv3874	CFP-10	10.8	4.3	✓		Dimer	94

^aSoluble aggregates.

CP, conserved protein.

M. smegmatis expression systems

Since the first report of introduction of foreign DNA into mycobacteria in 1987,¹⁹ M. smegmatis has been used as a model for pathogenic and slower-growing mycobacterial species. Isolation of an efficient plasmid-transformation mutant strain of M. smegmatis (mc²155) from the wild-type ATCC 607 strain²⁰ established an effective way of delivering DNA into this strain. Expression of foreign genes in mycobacteria was later made possible by generating shuttle vectors that used the M. bovis BCG hsp60 promoter for constitutive expression²¹ (e.g. pMV261 and pMV361). Inducible expression systems, on the other hand, provide a means of regulating gene expression and allow direct comparison of the induced and non-induced cells. Given the wide range of plasmids generated by the mycobacterial research community,²² a variety of inducible expression systems have been used to express and purify recombinant proteins from M. smegmatis host cells.

Acetamidase promoter-based expression system

This system makes use of the inducible acetamidase promoter of M. smegmatis.²³ Protein expression is induced by the addition of acetamide to cell cultures,²⁴ although the induction mechanism is not fully understood. The vector pJAM2 was developed by Triccas et al. in 1998²⁵ through cloning a 1.5 kb fragment of the acetamidase promoter of M. smegmatis into the pJEM12 shuttle vector.²⁶ This was the first mycobacterial expression vector possessing two important characteristics; the ability to regulate high-level expression of recombinant genes (acetamide induction) and a simple and efficient method to purify the recombinant proteins (through a His₆-tag). The pJAM2 vector is reported to be unstable for expression of proteins in Mtb, due to deletion of the promoter region.²⁷

A similar system was later developed to study the proteins present in the region of deletion 1 (RD1) of Mtb.²⁸ The pSD vector series generated in this study were constructed by inserting a 2.6 kb fragment of the M. smegmatis acetamidase promoter into the promoterless plasmid pMV206. Recombinant proteins could be expressed with His₆- or HA-tags for downstream analyses.²⁸

T7 promoter-based vectors

The T7 expression system for E. coli²⁹ provides the most common strategy used for heterologous protein production. In this system, the RNA polymerase from T7 bacteriophage transcribes the genes cloned downstream of the T7 promoter.²⁹ Exploiting the efficiency of this system in E. coli and its adaptability for use with other expression systems,³⁰ T7 promoter-based vectors³¹ (e.g. pYUB1049 and pYUB1062 shuttle vectors) have been developed for protein expression in M. smegmatis host cells. These vectors have the advantage that expression and solubility trials can be carried out in parallel in E. coli and M. smegmatis.

The pYUB1049 vector has been used as the basis for three further vectors; pDESTsmg,⁴ pYUB28b and pYUBDuet.³² pDESTsmg is a destination vector for the Gateway® cloning system with the capability of expressing recombinant proteins with an N-terminal His₆-tag.⁴ The pYUB28b vector harbors the multiple cloning site from the pET28b plasmid (Novagen) and can express proteins with N– and C–terminal His₆–tags.³² The pYUBDuet vector is a co-expression vector for simultaneous expression of two genes from two separate T7 promoters.³² The multiple cloning site for this vector comes from pETDuet-1 vector (Novagen), expressing one of the proteins with an N-terminal His₆-tag and the other one with a C-terminal S-tag.

Tetracycline-inducible expression system

This system has been widely used in both prokaryotic and eukaryotic organisms, and is capable of regulating gene expression both in vitro and in vivo.³³ Expression in this system is regulated through a constitutively expressed tetracycline repressor (TetR) protein, where TetR binding to the operator prevents transcription in the absence of inducer (e.g. anhydrotetracycline). Inducer binding to TetR, however, relieves the repression and transcription takes place.³⁴ Considering the ability of the tetracycline-inducible expression system to tightly regulate gene expression in mycobacteria,^35,36 it has been used in various contexts for conditional expression of essential genes in pathogenic mycobacteria.^33,37

A series of tetracycline-inducible Gateway®-compatible vectors have been reported that use tandem affinity purification tags (His₆-Strep-Strep) at either N- or C-termini of recombinant proteins.³⁸ Another vector toolbox (pST series) allows tetracycline-inducible expression of proteins with N-terminal (His₆-tag) and C-terminal (FLAG) tags.³⁹ The authors of this work have also generated a suicide plasmid for creating gene replacements, a plasmid for chromosomal integration, a plasmid for gene expression under hypoxic conditions and a vector for co-expression of two proteins under control of two independent promoters.³⁹

l-lactamase promoter-based expression system

Expression of recombinant proteins in this system is under the control of the β-lactamase promoter of Mycobacterium fortuitum.⁴⁰ This expression system has been mainly used to express and purify a variety of mycobacterial antigens from M. smegmatis.^41–43 The protein yield of the Mycobacterium leprae 35 kDa antigen was reported to be similar to that observed with the acetamidase-based expression system²⁵ but purification is based on antibody affinity chromatography.

Other expression systems

Alternative inducible expression systems for mycobacterial proteins include arabinose-inducible,⁴⁴ pristinamycin-inducible,⁴⁵ nitrile-responsive, and thiostrepton-responsive systems.⁴⁶ These expression systems provide a means of regulating gene expression in a wider range of mycobacterial hosts. The latter two expression systems have been reviewed elsewhere.⁴⁶

Co-expression systems

A number of methods for investigating protein-protein interactions in mycobacterial host cells have been reported, based on bacterial two-hybrid systems.⁴⁷ These methods include the mycobacterial protein fragment complementation (M-PFC)⁴⁸ and Split-Trp⁴⁹ systems. These do not, however, allow purification of the native protein complex due to the presence of fused reporter proteins.

In another approach, a tunable co-expression vector (pTetCoex) has been developed, based on the tetracycline-inducible expression system, to study protein-protein interactions in M. smegmatis cells.⁵⁰ This vector allows co-expression to be optimized by keeping the expression of one protein constant, while changing the expression level of its partner. Expression is monitored by antibody-based methods.⁵⁰

As described earlier, pYUBDuet³² and pST-2K³⁹ are co-expression vectors that can be used to over-express and purify stable protein complexes from M. smegmatis host cells. A recent report describes expression of a recombinant protein in M. smegmatis host cells, to be used as a bait for identification and purification of partner proteins that show weak or transient interactions, using in vivo formaldehyde crosslinking.³⁸ This method could potentially also be used with the aforementioned co-expression vectors, where the components of protein complex are over-expressed together.

M. smegmatis strains

Since the isolation of the M. smegmatis mc²155 strain,²⁰ it has been used as the workhorse of mycobacterial research, especially for expressing genes from mycobacterial species. It has been speculated that the M. smegmatis mc²155 strain has acquired changes in plasmid replication, enabling a high propensity for transformation.²⁰ Observed modifications in the composition and structure of the cell envelope may also contribute to explain this useful property.⁵¹ The exact mechanism by which this strain has acquired its high transformability, however, is not yet understood.⁵¹

The M. smegmatis mc²4517 strain is an expression strain that has been generated to make use of T7 promoter-based expression system.³¹ In this strain, a copy of T7 RNA polymerase is integrated into the chromosome under control of the acetamidase promoter from M. smegmatis.³¹ T7 RNA polymerase production is induced with acetamide and in turn transcribes the recombinant genes for subsequent translation. Expression can also be induced using lactose or IPTG in this system.⁵² This strain is used to express proteins using T7 promoter-based vectors described earlier.

When recombinant proteins containing a His₆-tag are expressed from mc²155 or mc²4517 strains, a major contaminant protein (the Hsp60 chaperone GroEL1) often co-purifies with the proteins.^4,53 This protein has a histidine-rich C-terminal tail, enabling it to bind to the purification beads. A mutant strain of M. smegmatis mc²155, so called GroEL1ΔC, has been generated to improve purification efficiency by removing the coding sequence for the histidine residues.⁵⁴

M. smegmatis advantages

The rationale behind using mycobacterial host cells for production of mycobacterial proteins is that these micro-organisms are more closely related and could therefore provide the specific requirements for proper folding, soluble expression and production of biologically and immunologically active proteins.

Molecular chaperones

The co-purification of M. smegmatis GroEL1 protein with recombinant proteins,^4,53 described above, implies that the presence of specific mycobacterial chaperones may facilitate the efficient folding and soluble expression of Mtb proteins. It is interesting to note that the Mtb homologs of GroEL are different from the E. coli counterparts, in that they are ATP independent and do not form the classical assemblies observed in E. coli.⁵⁵

Post-translational modifications

Post-translational modifications of proteins are more common in eukaryotes, although new technologies have revealed the presence of a diverse set of post-translational modifications in prokaryotes.⁵⁶ A wide range of post-translational modifications have been reported in the Mtb proteome,⁵⁷ and a clear advantage of mycobacterial host cells is their capability for post-translational modifications of the recombinant proteins.^28,58–61 The ability of mycobacterial expression systems to provide physiologically relevant post-translational modifications may in part contribute to their efficient production of biologically and immunologically active proteins.

Metabolites/ligands

Mycobacteria occupy a distant location on the evolutionary tree compared with E. coli, consistent with their unique environmental niches and physiological needs. This could, in principle, be associated with differences in biochemical pathways, enzymatic functions and metabolites in mycobacteria. For instance, a metabolome-based reaction network analysis revealed that reactions involving glutamate are central to the metabolism of Mtb and M. leprae, whereas pyruvate fills this role in the E. coli metabolome.⁶²

A great advantage of using M. smegmatis as an expression host is the presence of ligands that might not be present in E. coli. In this context, co-purification of a rare flavin derivative, coenzyme F₄₂₀, has been demonstrated upon over-expression and purification of an Mtb protein from M. smegmatis.⁶³ This coenzyme is sporadically present in microorganisms and not present in E. coli, likely explaining the unsuccessful attempts to express this as a soluble protein in E. coli.⁶³

Applications in structural biology

In macromolecular crystallography, single- and multi-wavelength anomalous diffraction (SAD/MAD) methods provide a powerful approach for experimental phasing of crystal structures.⁶⁴ Selenomethionine(SeMet)-substituted proteins provide a valuable method for the solution of protein crystal structures using SAD/MAD methods.⁶⁵ These proteins are prepared by replacing methionine residues with SeMet during protein expression. SeMet incorporation protocols have already been described for E. coli,⁶⁶ baculovirus,⁶⁷ mammalian,⁶⁸ and yeast⁶⁹ expression systems, as well as in vitro expression system.⁷⁰

Both T7 promoter-based⁵² and acetamide-based⁷¹ expressions systems in M. smegmatis have been successfully reported to prepare SeMet-substituted proteins for experimental phasing. This method has led to determination of the crystal structures of three proteins from Mtb by the use of MAD phasing; F₄₂₀-dependent glucose-6-phosphate dehydrogenase (FGD1),⁷² polyketide synthase 11 (PKS11),⁷¹ and Δ¹-pyrroline-5-carboxylic dehydrogenase (PruA).⁷³

A search of the PDB for entries with a reported alternative expression host for mycobacterial proteins reveals the presence of 40 structures, belonging to 12 unique proteins from Mtb (Fig. 2). Nine of these 12 proteins were produced in M. smegmatis and one in each of M. vaccae,¹⁸ Rhodococcus jostii,⁷⁴ and Pseudomonas putida⁷⁵ as the expression host (Fig. 2). Although the present list may not be exhaustive, it clearly demonstrates the successful application of M. smegmatis as a host for obtaining proteins in the quantity and high quality required for X-ray crystallography.

Figure 2.

Crystal structure of Mtb proteins expressed using alternative expression hosts. All proteins were expressed in M. smegmatis expression host, except for 1GN2 (M. vaccae), 4JN6 (Rhodococcus jostii), and 2VF2 (Pseudomonas putida).

Practical aspects

Given the different doubling time of M. smegmatis cells (∼3 h) compared with that of E. coli (∼20 min) and the waxy nature of the mycobacterial cell wall, growth conditions, and protocols should be accordingly adapted for M. smegmatis cultures.⁷⁶ Once transformed, it takes ∼3 days for colonies to appear on plates, and starting from single colonies it would take 2–3 days for 5–10 mL starter cultures to grow. Expression cultures are inoculated at a dilution of 1:100 using starter cultures and expression is usually carried out for 3–4 days. The cultures must also be supplemented with detergents (e.g. Tween80 or tyloxapol) to prevent cell clumping.

The protein yield obtained from M. smegmatis cultures varies significantly, as for other expression systems. In the author's lab, ∼10 mg protein per liter of cell culture is achieved for some proteins⁵² while a few others are expressed as insoluble proteins. Our early results on using M. smegmatis as a rescue expression host indicated that >60% of proteins that formed inclusion bodies when expressed in E. coli are soluble when expressed in M. smegmatis.⁴ Since then, we have successfully used M. smegmatis to express and purify over 25 mycobacterial proteins, leading to crystallization and determination of the crystal structures for eight of these proteins. These results support the use of M. smegmatis host cells as a routine strategy for protein production rather than just a rescue/salvage one (Table 1).

Expression of recombinant proteins in M. smegmatis sometimes results in slow growth, very low or no expression or insoluble expression. These symptoms may be due to toxicity of these particular proteins, imposing heavy physiological burden on to the cells. A number of methods could be used to resolve this issue, including using plasmids with lower copy numbers,²² use of weaker promoters with lower expression level⁷⁷ or using a promoter that is more tightly regulated.³⁹ Alternatively, other expression hosts that are described in this manuscript could be of use. In addition, some proteins that are part of hetero-oligomeric protein complexes might not be expressed as individual proteins, in which case co-expression with the partner protein could rectify the expression problem.⁵⁰

Concluding Remarks

Genetic manipulation of mycobacterial species has gone through an extensive progression over the last 30 years, leading to a better understanding of the function, structure and regulation of complex biological mechanisms in mycobacteria. Determination of the genome sequences for two major pathogens, Mtb⁵ and M. leprae,⁷⁸ has since opened up a new window to investigate the biology and pathogenesis of these micro-organisms through a wide range of methods. The ultimate aim of the studies is to clear the way for novel and better therapeutics and diagnostics through studying the structure and function of mycobacterial proteins. Using M. smegmatis as an expression host can enable researchers to tackle proteins that were not amenable before and thereby produce biologically and immunologically active proteins for research and therapeutic applications.

Glossary

MAD

multi-wavelength anomalous diffraction

M-PFC

mycobacterial protein fragment complementation

Mtb

Mycobacterium tuberculosis

MTBSGC

M. tuberculosis Structural Genomics Consortium

PDB

Protein Data Bank

SAD

single-wavelength anomalous diffraction

SeMet

selenomethionine

TetR

tetracycline repressor