Regulation of Expression of mas and fadD28, Two Genes Involved in Production of Dimycocerosyl Phthiocerol, a Virulence Factor of Mycobacterium tuberculosis

Abstract

Transcriptional regulation of genes involved in the biosynthesis of cell wall lipids of Mycobacterium tuberculosis is poorly understood. The gene encoding mycocerosic acid synthase (mas) and fadD28, an adjoining acyl coenzyme A synthase gene, involved in the production of a virulence factor, dimycocerosyl phthiocerol, were cloned from Mycobacterium bovis BCG, and their promoters were analyzed. The putative promoters were fused to the xylE reporter gene, and its expression was measured in Escherichia coli, Mycobacterium smegmatis, and M. bovis BCG. In E. coli, the fadD28 promoter was not functional but the mas promoter was functional. Both fadD28 and mas promoters were functional in M. smegmatis, at approximately two- and sixfold-higher levels, respectively, than the BCG hsp60 promoter. In M. bovis BCG, the fadD28 and mas promoters were functional at three- and fivefold-higher levels, respectively, than the hsp60 promoter. Primer extension analyses identified transcriptional start points 60 and 182 bp upstream of the translational start codons of fadD28 and mas, respectively. Both promoters contain sequences similar to the canonical −10 and −35 hexamers recognized by the σ⁷⁰ subunit of RNA polymerase. Deletions of the upstream regions of both genes indicated that 324 bp of the fadD28 and 228 bp of the mas were essential for promoter activity. Further analysis of the mas promoter showed that a 213-bp region 581 bp upstream of the mas promoter acted as a putative transcriptional enhancer, promoting high-level expression of the mas gene when present in either direction. This represents the identification of a rare example of an enhancer-like element in mycobacteria.

Mycobacteria are important human pathogens responsible for major life-threatening diseases such as tuberculosis and leprosy. One-third of the world population have latent Mycobacterium tuberculosis infection, and 5 to 10% of them develop active tuberculosis at some time during their lives (38). Currently, 8 million new cases of active tuberculosis are diagnosed each year, with 2 million deaths. The increased incidence of tuberculosis as an AIDS-related illness, together with the burgeoning problem of multidrug-resistant mycobacterial strains, has established tuberculosis as a disease of major public health significance (39). Furthermore, the Centers for Disease Control and Prevention has classified multidrug-resistant M. tuberculosis as a class C organism with potential for use in creating major public health problems in population centers.

The very high lipid content (50 to 60%) of the mycobacterial cell wall makes it an effective permeability barrier both to the host's own immune system and to antimycobacterial therapeutic agents. Some of these lipids, such as dimycocerosyl phthiocerol (DIM), are necessary for virulence (6, 8,19). In recent years, considerable progress has been made in identifying genes involved in the biosynthesis of some of the unique lipids of mycobacterial cell wall. Examples include the mycocerosic acid synthase gene (mas) (24), an associated acyl coenzyme A (CoA) synthase gene (fadD28) (13) involved in the esterification of the acid to phthiocerol to generate DIM, some of the genes involved in the production of phthiocerol (2, 6, 8), and several mas-like genes responsible for the synthesis of some of the multiple methyl-branched fatty acids that are esterified to trehalose and sulfated trehalose (12, 33). However, little is known about the regulation of expression of these genes. An understanding of the mechanisms governing the regulation of gene expression in pathogenic mycobacteria can be very helpful in the development of novel therapeutic approaches to combat this pathogen. Recently, a number of mycobacterial promoters have been identified in fast- and slow-growing species. A few of these promoters have elements typical of Escherichia coli σ⁷⁰ consensus promoters (25, 26, 36, 37), but the vast majority have GC contents reflective of the mycobacterial genome (65 to 70%) and are poorly expressed in heterologous hosts (3, 7, 28, 32).

In this paper we report the results of analysis of the promoters of mas and fadD28, which are involved in the synthesis of a virulence factor, DIM. Our results indicate that these promoters are stronger than the well-characterized hsp60 promoter (31). We show that there is an enhancer-like element about 580 bp upstream from the basic mas promoter. These promoters, being stronger than previously characterized mycobacterial promoters, may be useful in expressing antigens in Mycobacterium bovis BCG for use in recombinant vaccines.

MATERIALS AND METHODS

Bacterial strains and plasmids.

Descriptions of the bacterial strains and plasmids used in this study have been presented previously (2, 9, 15).

General cloning procedures.

All restriction endonuclease digestion and ligation reactions were performed as outlined elsewhere (29). DNA fragments used for cloning and labeling studies were eluted from low-melting-point agarose, according to the manufacturer's guidelines (FMC, Newark, Calif.). Recombinant plasmids were transformed into E. coli DH5α and selected on Luria-Bertani plates containing 50 μg of kanamycin per ml. Mycobacterium smegmatis and M. bovis BCG transformants, containing pRCX3 (9) derivatives, were selected on Middlebrook 7H11 plates (Difco) supplemented with oleic acid-albumin-dextrose complex (OADC; Difco) and kanamycin (25 and 12.5 μg per ml, respectively).

PCR.

Specific fragments were amplified by PCR with the primers described in Table 1 and Pfu polymerase (Stratagene, La Jolla, Calif.) according to the manufacturer's recommendations, with dimethyl sulfoxide being included in the reaction mixtures at a final concentration of 5% (28 sequential reaction cycles at 94°C for 1 min, 58°C for 45 s, and 72°C for 1 min). Reactions were completed with an incubation at 72°C for 10 min.

TABLE 1.

Oligonucleotide primers used in this study

Oligonucleotide	DNA sequence (5′-3′)
A1	GCAGGCACGCACACGCGCGCA
A2	ACGGGGACTACGACAACTCCA
A3	AGCTGGTCAGATTCGACTTGG
A4	CCGGCGGGGGGGCGGCAGCAC
A5	GGGCTGTTCGACCAGCGGATT
M	ACCGCAACGGGAGTGAC
M1	GGATCCGGGTTGGCTCG
M2	CCAGGTCCGAGGTGAGA
M3	CACGATCACCGATGACC
M4	AGTCGCATCTGGCCAGC
M5	TTGCAGACGCGCACACG
M6	CAGCGTTCGCCGATACA
M7	TGCAACGCGCCGAGAAA
M8	CCCCGGGCCGCCGCGCA
M9	CGCCACATCACATCCGA
EN-F	AGCTCCTGTGCCACATT
EN-R	CACACCCCGACGGCGAC
xy1E	CTTGATCGCGGTGCTCTCG
PEfadD28-1	GCAACGCAGCGGGAAGGGAACGGACACTC
PEfadD28-2	GTTCGTAATCCATAAACGTGAAGGCCGGGT
PEmas-1	GTGTTACCCGACGACATCAGAGAATTTTCA
PEmas-2	GCTCTCCCAGCTCTTAAGTAATCCGAGCCA

Construction of deletion derivatives.

DNA fragments containing deletion derivatives of the fadD28 and mas promoters were synthesized by PCR amplification. Oligonucleotides (Table 1) were designed to generate either a PstI (for fadD28) or HindIII (for mas) site at the end of the fragments. In each case, the template used for amplification was a genomic M. bovis BCG DNA segment containing residues 1 to 1770 of the sequence with accession no. M95808 (24), cloned in the EcoRI site of pUC19; this segment contains the shared 650-bp region which includes both the fadD28 and mas promoters. The deletion derivatives were generated by using the following primer pairs: for the fadD28 promoter, primers A1 and A2 (96-bp fragment), A1 and A3 (146-bp fragment), A1 and A4 (222-bp fragment), and A1 and A5 (378-bp fragment); for the mas promoter, primers M and M1 (126-bp fragment), M and M2 (228-bp fragment), M and M3 (360-bp fragment), M and M4 (542-bp fragment), M and M5 (669-bp fragment), M and M6 (768-bp fragment), M and M7 (870-bp fragment), M and M8 (1023-bp fragment), M8 and M9 (560-bp middle fragment), and EN-F and EN-R (213-bp enhancer fragment). fadD28 and mas deletion derivatives were constructed by cloning the corresponding DNA fragments into the PstI-digested (for fadD28 fragments) and HindIII-digested (for mas fragments) pRCX3 (9), a generous gift from V. Deretic.

RNA isolation.

Plasmid pTSP1 was generated by cloning 1,770 bp of the M. bovis BCG Tice genomic DNA directly upstream of the mas translational start codon and the first 30 bp of the coding mas sequence in the HindIII site of pRCX3. Mid-log-phase M. smegmatis cells, containing plasmid pTSP1, were harvested from a 100-ml culture by centrifugation and resuspended in 2 ml of ice-cold 50 mM Tris-HCl, pH 7.5. The cell suspension was transferred to a polypropylene screw-cap tube containing 500 mg of 425- to 600-mm acid-washed glass beads and homogenized at maximum speed for 1 min in a minivortexer (VWR, Chicago, Ill.). To each tube 200 μl of 20% sodium dodecyl sulfate was added, and the mixture was incubated at 67°C for 1 min. The lysate was transferred to a QIA shredder unit (Qiagen, Chatsworth, Calif.), centrifuged to degrade chromosomal DNA, and applied to a Qiagen Total RNA-tip 100. Completion of total RNA isolation was performed as outlined in the Qiagen Total RNA handbook.

Primer extension analysis.

Two sets of radiolabeled oligonucleotide primers (PEfadD28-1 and PEfadD28-2; PEmas-1 and PEmas-2) that were complementary to regions downstream of fadD28 and mas promoters, respectively, were hybridized separately to total RNA isolated from exponentially growing M. smegmatis cells harboring plasmid pTSP1. Hybridization was performed at 85°C for 10 min, followed by annealing at either 45 or 65°C for 3.5 h and extension with reverse transcriptase at 42°C for 1.5 h (Superscript RNase H⁻ reverse transcriptase; Life Technologies). A DNA sequence ladder of dideoxy sequencing products was generated by using each oligonucleotide and plasmid pTSP1 as a template.

Electroporation of M. smegmatis and M. bovis BCG.

M. smegmatis and M. bovis BCG cultures were grown to mid-log phase in Middlebrook 7H9 broth supplemented with 0.5% Tween 80 and albumin-dextrose complex (Difco). Cultures were treated with glycine (1.5%) for 24 h prior to harvest, and 100- to 200-μl aliquots were used in transformation experiments using previously outlined parameters (15).

Assay of CDO activity.

Catechol 2,3-dioxygenase (CDO) activity was determined as previously described (9). Briefly, E. coli harboring the promoter constructs grown on Luria-Bertani agar with kanamycin and mycobacterial colonies grown on Middlebrook-OADC-kanamycin plates were sprayed with an aqueous solution of 100 mM catechol in 50 mM potassium phosphate buffer, pH 7.5. Colonies expressing xylE were identified by their yellow color. For quantitative analysis, confluent cultures were harvested from agar plates with 0.85% saline, centrifuged and washed in the same buffer, resuspended in 50 mM potassium phosphate buffer (pH 7.5) containing 10% acetone, and subjected to three 10-s sonications at 4°C with the needle probe of a Sonifier cell disrupter with a double-step microtip assembly (Branson Ultrasonic; UWR Scientific Products). After the cell debris had been removed by centrifugation, CDO activity in the cell extracts was determined by monitoring the increase in absorbance at 375 nm over a 2-min period in the presence of 0.33 mM catechol. The experimental values and standard errors were calculated from the results of three experiments, each with four independent cultures containing the specific construct. Protein concentrations were determined by the Bradford method (5) with bovine serum albumin as the standard. One unit of CDO activity is defined as the amount of enzyme that oxidizes 1 μmol of catechol min⁻¹ at 24°C.

RESULTS

Previous studies reported the cloning and characterization of the genes encoding an acyl-CoA synthase (fadD28) (13) and mycocerosic acid synthase (mas) (24). These two genes, involved in the synthesis and utilization of mycocerosic acids, are located adjacent to each other in the M. bovis BCG, M. tuberculosis, and Mycobacterium leprae genome but are transcribed in opposite directions. In order to explore the mechanisms governing expression of the genes involved in mycocerosic acid synthesis, the promoter regions of the mas and fadD28 genes were fused to the xylE reporter gene in the promoter-probe vector pRCX3. The xylE gene product CDO converts catechol into 2-hydroxymuconic semialdehyde. This compound has a bright yellow color with an absorbance maximum at 375 nm. Promoter activity was determined by measurement of CDO activity in cell extracts.

Analysis of fadD28 and mas promoter expression in E. coli, M. smegmatis, and M. bovis BCG.

A 1,055-bp region (nucleotides −1025 to +30 of the mas gene) containing the fadD28 and mas promoters was amplified by PCR and cloned in both orientations into pRCX3, giving rise to pmas8(+) and pmas8(−). Constructs in which the fadD28 and mas promoters were transcribed in the same direction as the xylE gene were selected by PCR amplification with either of the original two primers (M and M8) (Table 1) used to generate the cloned fragment and the primer xylE (Table 1), specific for the region adjacent to the cloned fragment. Since fadD28 and mas are transcribed in opposite directions in pmas8(+), xylE is in the same orientation as mas, and thus, its expression would be driven by the mas promoter. In pmas8(−), xylE expression would be driven by the fadD28 promoter. Plasmids pmas8(+) and pmas8(−) were transformed into E. coli, M. smegmatis, and M. bovis BCG, and CDO activities were determined in cell extracts. E. coli carrying plasmid pmas8(−) failed to show any significant CDO activity, indicating that the fadD28 promoter was not expressed in E. coli. However, pmas8(+) resulted in a significant CDO activity of approximately 500 mU/mg of protein in E. coli, indicating that the mas promoter is functional in this organism.

Since the activities of mycobacterial promoters are likely to be manifested more accurately in mycobacterial hosts, we tested fadD28 and mas promoter derivatives in fast-growing M. smegmatis and slow-growing M. bovis BCG. In M. smegmatis, a fadD28 promoter driving xylE gene expression [pmas8(−)] gave approximately 2,700 mU of CDO activity per mg of protein. The level of expression of xylE in M. bovis BCG was comparable to that seen in M. smegmatis (Table 2). The mas promoter-driven expression of xylE [pmas8(+)] was robust in both M. smegmatis and M. bovis BCG. In this case, expression in M. bovis BCG was slightly (20%) less than that observed in M. smegmatis (5,700 mU/mg of protein compared to 7,300 mU/mg of protein). The promoters of both fadD28 and mas were considerably stronger than the hsp60 promoter in driving the expression of xylE. The fadD28 promoter showed two- to threefold-higher activity than the hsp60 promoter in M. smegmatis and M. bovis BCG. The mas promoter showed five- to sixfold-higher activity than the hsp60 promoter in mycobacteria (Table 2).

TABLE 2.

Comparison of reporter gene (xylE) expression driven by fadD28, mas, and hsp60 promoters

Promotera	xylE (CDO) activity (mU/mg of protein)
	M. smegmatis	M. bovis BCG
hsp60	1,300 ± 127	1,100 ± 78
hsp60 (reverse)	122 ± 37	100 ± 12
fadD28	2,667 ± 193	2,978 ± 211
fadD28 (reverse)	166 ± 15	156 ± 26
mas	7,288 ± 478	5,700 ± 312
mas (reverse)	3,078 ± 268	2,900 ± 127

^a“Reverse” indicates that the orientation of the promoters is opposite to the direction of transcription of xylE. The high activity observed with the 1-kb mas promoter in the reverse orientation is due to the activity of the fadD28 promoter. In the case of the fadD28 promoter, only a 378-bp segment that does not include the mas promoter is used.

Mapping of the 5′ end of the fadD28 and mas transcripts.

The transcription start points of fadD28 and mas genes were determined by primer extension analysis, using total RNA isolated from exponentially growing M. smegmatis cells containing plasmid pTSP1. A radiolabeled oligonucleotide, designated PEfadD28-1 (Table 1), complementary to nucleotides 37 to 8 of the fadD28 gene, was hybridized to total RNA and extended by using reverse transcriptase. The size of the cDNA product, corresponding to the distance from the 5′ end of the transcript to the 5′ end of the primer, was determined by comparison to the electrophoretic mobility of a DNA sequence ladder, generated with the same oligonucleotide and plasmid pTSP1 DNA. A major transcription start point mapped to a C residue on the nontranscribed strand, 60 bp upstream of the fadD28 translation initiation codon (Fig. 1A). The location of this transcription start point was verified by using a second oligonucleotide, PEfadD28-2 (complementary to oligonucleotides 95 to 65 of the fadD28 gene) (data not shown). A potential Pribnow (−10) box is present at a suboptimal 7 bp upstream of the fadD28 transcriptional start point. A putative −35 region (GTCGTC) is present at the optimal 17 bp upstream of the −10 region; this sequence (TAGGTT) matches the σ⁷⁰ consensus sequence in three of six positions.

FIG. 1.

Mapping of the transcription initiation sites of the fadD28 (A) and mas (B) genes by primer extension analysis. Lanes G, A, T, and C, DNA sequence ladder of the fadD28 and mas promoter regions; lanes 1 and 2, primer extension with annealing temperatures of 45 and 65°C, respectively. The major extension products are indicated, and the transcription initiation sites of both genes are shown by arrows (details are in Materials and Methods).

A radiolabeled oligonucleotide, designated PEmas-1 (Table 1), which was complementary to residues 30 to 60 of the mas gene, was hybridized to total RNA and used to determine the start point of mas transcription. A major hybridization band mapped to a T residue, 182 bp upstream of the translation initiation codon of mas (Fig. 1B). The location of the mas transcription start point was also confirmed by using a second radiolabeled oligonucleotide, PEmas2, complementary to nucleotides 88 to 117 of the mas gene (data not shown).

A putative Pribnow (−10) hexamer (TGTGCT) was detected 10 bp upstream of the mas transcription start point, which matched the canonical σ⁷⁰ −10 sequence at positions 1, 3, and 6. A putative −35 element (GTGAGA), located at a suboptimal 16 bp upstream of the −10 hexamer, was also present and corresponded to the −35 element of σ⁷⁰ at positions 2, 3, 4, and 6. Both the −10 and −35 mas hexamer sequences have undergone GC substitution, which is reflected in the fact that the promoter region (positions −1 to −50 upstream of the mas transcription start point) displays a GC content of approximately 56%. The average GC contents of M. tuberculosis and M. smegmatis promoters are approximately 57 and 43%, respectively (3).

Deletion analysis of the fadD28 and mas promoter regions.

In order to define the regions essential for promoter activity, a series of deletion derivatives of the fadD28 5′ upstream region were created in pRCX3. Fragments of 96, 146, 222, and 378 bp were amplified by PCR and cloned in both orientations into the PstI site of pRCX3, creating plasmids pfadD28-1(+) and pfadD28-1(−), pfadD28-2(+) and pfadD28-2(−), pfadD28-3(+) and pfadD28-3(−), and pfadD28-4(+) and pfadD28-4(−), respectively (Fig. 2). In M. smegmatis, plasmid pfadD28-4(+) resulted in a CDO activity of approximately 2,700 mU/mg of protein, which was comparable to that expressed in pmas8(−)-containing cells. This result suggests that the additional 622 bp present in pmas8(−) does not play a significant role in expression of the fadD28 promoter and that elements essential for fadD28 transcription are located in the 324-bp region directly upstream of the fadD28 translation initiation codon. Deletion of an additional 156 bp from the 5′ end of the fadD28 promoter region [pfadD28-3(+)] resulted in a reduction in CDO activity by approximately 50% relative to levels expressed in pfadD28-4(+)-containing cells (Fig. 2). The sequence of the promoter element in this construct contains intact −10 and −35 elements; however, the reduction in xylE expression suggests that further elements essential for full promoter activity exist between 168 and 324 bp upstream of the fadD28 translation initiation codon. When the −35 region was deleted [pfadD28-2(+)], CDO activity was reduced by approximately 75% of the level detected when the entire promoter region was present. Therefore, the −35 region contributes significantly to but is not absolutely required for fadD28 promoter function. Deletion of both the −10 and −35 regions [pfadD28-1(+)] completely abolished fadD28 promoter activity. Constructs containing each of the fadD28 promoter deletion derivatives cloned in the opposite orientation to the xylE gene did not display CDO activities above background levels, indicating the absence of a functional promoter in the direction opposite to fadD28 gene transcription (Fig. 2).

FIG. 2.

Reporter gene expression driven by fadD28 and its derivatives fused to the xylE reporter gene. The locations of the ATG translation initiation codons of fadD28 and mas are marked. CDO activities in M. smegmatis and M. bovis BCG are expressed in milliunits per milligram of protein in extracts from cells at mid-exponential phase. +, promoter in the same direction as xylE; −, promoter in the opposite direction.

In the case of the mas promoter, 126-, 228-, 360-, 542-, and 1,023-bp segments directly upstream of the translation initiation codon were cloned in both orientations into the HindIII site of pRCX3, generating plasmids pmas-1(+) and pmas-1(−), pmas-2(+) and pmas-2(−), pmas-3(+) and pmas-3(−), pmas-4(+) and pmas-4(−), and pmas-8(+) and pmas-8(−), respectively (Fig. 3). pmas-2(+), pmas-3(+), and pmas-4(+) resulted in CDO activities of approximately 2,000 mU/mg of protein in M. smegmatis; these results indicate that the 228-bp region upstream of the mas start codon is sufficient to promote expression of the xylE gene. These findings are in agreement with the results of primer extension analysis experiments, which indicate that this region contains intact −10 and −35 sequence recognition elements.

FIG. 3.

xylE gene expression driven by mas promoter and its derivatives fused to xylE gene. The translation initiation sites are marked and the enzymatic activity of the xylE gene product (CDO) is shown in milliunits per milligram of protein in the cell extracts. +, promoter in the same direction as xylE; −, promoter in the opposite direction.

CDO levels exhibited by M. smegmatis containing pmas-2(+), pmas-3(+), or pmas-4(+) were approximately threefold lower than those expressed in cells harboring pmas-8(+) (Fig. 3). These results suggest one of two possibilities: either a second promoter is functional in the direction of mas transcription, or mas promoter activity is augmented by an upstream activator sequence, itself incapable of promoting gene expression. To differentiate between these two possibilities, additional constructs were made. The 560-bp region present in pmas8(+) but absent from pmas4(+) (nucleotides −1023 to −463) was cloned in both orientations into pRCX3, generating plasmid pmas-9(+) and pmas-9(−) (Fig. 3). In M. smegmatis, pmas9(+) failed to exhibit significant CDO activity, thereby excluding the possibility that a mas-independent promoter is present in this region and functional in the direction of mas transcription. This DNA segment in the reverse orientation showed promoter activity reflecting the presence of the fadD28 promoter.

In order to localize the putative upstream activator element, three additional constructs containing 669, 768, and 870 bp upstream of the mas initiation codon were created [plasmids pmas-5(+) and pmas-5(−), pmas-6(+) and pmas-6(−), and pmas-7(+) and pmas-7(−)] (Fig. 3). Optimal manifestation of the mas promoter activity occurred only in M. smegmatis cells, containing at least 768 bp of the mas upstream region [either pmas-6(+) or pmas-7(+)]. Therefore, a segment which spans nucleotides −764 to −670 should contain the putative activator sequence; this region was amplified and cloned in both orientations into the BclI site of pmas3(+), creating plasmids pmas-3(+)(EN+) and pmas-3(+)(EN−) (Fig. 3). In M. smegmatis, the presence of this potential enhancer region in either orientation elevated CDO activity to the pmas-8(+) levels. Therefore, an upstream activator sequence is essential for high level expression of the mas promoter.

When xylE expression was monitored in M. bovis BCG cells containing pmas8(+), pmas3(+), pmas3(+)(EN+), and pmas3(+)(EN−), the profile of mas promoter expression was similar to that in M. smegmatis; high-level mas expression showed an absolute dependence on the presence of the upstream activator sequence (Fig. 3). This element was equally effective when placed in either direction.

DISCUSSION

Transcriptional regulation of genes that play critical roles in virulence not only could provide better understanding of the process of pathogenesis but also could lead to new ways to prevent or control mycobacterial infection. The unique cell wall lipids of pathogenic mycobacteria have been thought to play a significant role in the ability of this pathogen to evade host defenses. In recent years, DIM has been identified as a cell wall lipid involved in virulence (6, 8, 19). Regulation of synthesis of this lipid involves expression of mas and an associated acyl-CoA synthase gene, fadD28. Therefore, we sought an understanding of the regulation of expression of these genes by examining their promoters. Our results show that the fadD28 gene promoter does not function in E. coli but that the mas promoter is functional in E. coli. Although the mas promoter is functional in E. coli, its activity is only 7 to 10% of that in mycobacterial species. This is in keeping with other studies which show that a number of M. smegmatis promoters support 2- to 100-times-lower chloramphenicol acetyltransferase activities in E. coli than in mycobacteria (10). A limited number of mycobacterial promoters have been shown to be functional in E. coli, including the M. bovis BCG hsp60 promoter (36), the β-galactosidase promoter from Mycobacterium paratuberculosis (26), the Mycobacterium fortuitum β-lactamase promoter (37), and the P1 and P2 promoters of the mycobacteriophage repressor L5 (27). However, it has been reported that only 12% of M. smegmatis promoters are active in E. coli (10), and 80% of promoters from the closely related Streptomyces spp. are nonfunctional in E. coli (35). Low-level expression of promoters from mycobacterial species in E. coli has been attributed to differences in the principal vegetative sigma factor present in these two organisms. Whereas the −10 binding domains of RpoD from E. coli, MysA from M. smegmatis, and HrdB from Streptomyces are similar, the −35 binding region of the sigma factors of E. coli is very different from the corresponding −35 binding domains of the mycobacterial and streptomycete polymerases (3).

In order to assess the relative strength of mycobacterial promoters involved in the synthesis of multimethyl branched fatty acids, CDO activities expressed under the control of the fadD28 and mas promoters were compared to that expressed from the M. bovis BCG hsp60 promoter. The latter promoter controls expression of one of the major stress proteins in mycobacteria and has been used to drive expression of foreign antigens in BCG (34). The fadD28 and mas promoters showed two- to threefold and five- to sixfold-higher levels of expression of the xylE reporter gene, respectively, than the hsp60 promoter in mycobacterial hosts. Expression of high levels of antigens and immunity-boosting ingredients could be highly useful for the production of recombinant BCG vaccines and serodiagnostic reagents.

The transcription initiation site of fadD28 is located 60 bp upstream from the translational start site found in E. coli fadD. Even though it appears that M. smegmatis and M. tuberculosis gene transcription initiates more often at a purine than a pyrimidine (3), transcription initiation of both fadD28 and mas genes is at pyrimidines. mas transcription starts 182 bp upstream of the translational initiation site, which is farther upstream than the initiation sites of many of the other mycobacterial genes (1, 14, 37).

A notable feature of the Pribnow (−10) box of fadD28 is that it contains the highly conserved bases T, A, and T at positions 1, 2, and 6, with the third and fourth bases having undergone GC substitution, a general feature of the −10 region of mycobacterial promoters (3). The −35 element of fadD28 displays limited homology to the −35 regions of some σ⁷⁰-type promoters, including the M. bovis BCG hsp60 gene (36), the 16S rRNA genes of M. leprae and M. tuberculosis (16, 30), and the 85A antigen gene of M. tuberculosis (20). In each case, the −35 region is identical at positions 4 and 5 of the hexamer, the most conserved bases in this recognition sequence. An additional recognition sequence is present at positions −13 to −29 upstream of the fadD28 transcriptional start point (CAGCTTAGACTGGTCGA); this sequence matches the consensus binding sequence of the transcriptional regulator, FadR, in 11 of 17 positions. In particular, it contains the sequence CTGGT (−22 to −26), which has been defined as part of the FAD R operator in both the fadB and FadL genes (11). In E. coli, it has been shown that binding of FadR to this sequence on the acyl-CoA synthase gene, fadD, overlaps the −10 recognition hexamer and blocks transcription initiation by RNA polymerase. Addition of the inducer oleate results in a twofold increase in FAD D activity (4). However, preliminary results suggest that oleate addition fails to stimulate M. bovis BCG fadD28 activity (data not shown). A second FadR binding site, with a lower affinity for the regulatory protein, was also identified at positions −115 to −98 upstream of the fadD transcriptional start point (4). However, a corresponding regulatory element is not evident in the M. bovis BCG fadD28 promoter. For fadD28 expression, the −35 region does not appear to be essential, although it contributes to its expression. Previous studies also report that the −35 element of a number of different genes is not essential for maximal promoter activity (3). However, deletion or alteration of the −10 hexamer usually has detrimental consequences for transcriptional activity (18, 21).

The finding of an enhancer-like element that is fully functional when placed in either direction upstream of mas promoter was unexpected, as such elements are rare in bacterial promoters (17, 22). Upstream activator regions have been identified in mycobacterial genes. For example, in the katG promoter of M. tuberculosis, an upstream activator region was found 448 to 559 bp upstream from the translation initiation codon. The distance between this region and the promoter sequences does not appear to be critical for its activating effect (25). In the iniBAC promoter of M. tuberculosis, an upstream region 147 to 169 bp from the translational start site was found to be necessary for induction of this gene by antibiotics, and deletion of this region virtually abolished not only induced but also uninduced expression of the genes (1). The narK2 promoter of M. bovis BCG showed a requirement for the 133- to 222-bp segment upstream from the transcription start site for basal and dormancy-inducible reporter gene expression (14). In these cases, activity of such regions placed in the opposite orientation has not been tested, and therefore, it is not known whether they are real enhancer-like elements.

Upregulation of mas and fadD28 genes following phagocytosis of M. bovis BCG by the macrophage-like THP-1 cells has been observed (23). Whether the regulatory elements found in the mas and fadD28 genes in the present study are involved in this upregulation remains unknown. However, in view of the recent findings that DIM, the cell wall lipid whose production involves the enzymes encoded by these genes, is a virulence factor, it is possible that this lipid synthesis may be induced in the host macrophage. If such induction assists the pathogen to survive and/or multiply in the host, the transcriptional regulation of synthesis of such genes could provide novel targets for antimycobacterial therapy.