Single Nucleotide Polymorphisms That Cause Structural Changes in the Cyclic AMP Receptor Protein Transcriptional Regulator of the Tuberculosis Vaccine Strain Mycobacterium bovis BCG Alter Global Gene Expression without Attenuating Growth

Strains and plasmids.The strain used for all of the experiments was the previously described mutant of M. tuberculosis, H37Rv, deleted for most of the Rv3676 gene (23). This deletion was complemented with the pRB132 plasmid, described previously (23), carrying the M. tuberculosis Rv3676 gene (referred to here as pcrp_Mtb) transcribed from its own promoter or mutated versions carrying the BCG Rv3676 (pcrp_BCG) sequence. This plasmid was a derivative of pKP186, an integrase-negative derivative of the integrating vector pMV306 (28), kindly provided by K. G. Papavinasasundaram. The pKP186 plasmid derivatives were cotransformed along with the plasmid pBS-int carrying the integrase gene necessary to achieve integration of the plasmid into the chromosome. pBS-int lacks a mycobacterial origin of replication and is therefore lost from the bacterium. The plasmid pKP186 integrates into the single L5attB chromosomal site as a single copy by site-specific recombination (18).

In vitro growth determination.Bacteria were grown in 100 ml of Dubos broth containing 0.05% (vol/vol) Tween supplemented with 0.2% (vol/vol) glycerol and 4% (vol/vol) Dubos medium albumin in 1-liter polycarbonate culture bottles (Techmate) in a Bellco roll-in incubator (at 2 rpm) at 37°C. Optical density readings of aliquots removed were taken at 600 nm.

Growth of bacteria in mouse macrophage cell culture.Bone marrow cells extracted from the hind legs of 8-week-old adult female BALB/c mice were grown and infected with mycobacteria as described previously (23).

Growth of bacteria in the mouse model of TB infection.Strains of M. tuberculosis H37Rv were grown as rolling cultures in Dubos broth (as described above) to mid-exponential phase. Each strain was diluted in phosphate-buffered saline to give a suspension of approximately 10⁶ CFU ml⁻¹, and 0.2 ml of these suspensions was inoculated intravenously into 6- to 8-week-old female BALB/c mice. The infection was monitored by removing the lungs and spleens of three to five infected mice at intervals. The tissues were homogenized by shaking with 2-mm-diameter glass beads in chilled saline with a Ribolyser. Serial 10-fold dilutions of the resultant suspensions were plated onto Dubos 7H11 agar with Dubos oleic albumin complex supplement (Difco Laboratories, Surrey, United Kingdom). The numbers of CFU were determined after the plates had been incubated at 37°C for approximately 20 to 35 days.

Construction of mutated complementing plasmids for Rv3676.A Stratagene QuikChange XL site-directed mutagenesis kit was used with the complementing plasmid pRB132, described previously (23), which complements an Rv3676 deletion mutant strain of M. tuberculosis, H37Rv. To generate the helix-turn-helix (HTH) and cyclic nucleotide-binding (cNMP) mutant, the HTH mutagenesis was performed, followed by the NMP; the primers used were as described previously (27). The resulting plasmids were sequenced to confirm that the point mutations had been introduced and that no other bases had been changed in the gene sequence.

Protein structure modeling.A computer model of the mycobacterial cAMP receptor protein Rv3676 was built on the X-ray structure of the CRP from Escherichia coli (the catabolite gene activator protein [CAP]) solved to a resolution of 2.5 Å (21) (Protein Data Bank, code 1J59) (29). Protein modeling software QUANTA (release version 4.0; Accelrys) running on a Silicon Graphics INDIGO2 workstation under the UNIX operating system was used to build 30 models of the mycobacterial CRP. The different models were calculated by varying the initial model and optimizing an objective function using conjugate gradients and molecular dynamics with simulated annealing that employed the CHARMM force field (6). The model with the lowest objective function was chosen for further study, allowing the prediction of sites for mutation and the explanation of the effect of the SNPs on protein function.

Transcription analysis using DNA microarrays.Bacteria were grown in Dubos medium (supplemented with 0.2% [vol/vol] glycerol, 4% [vol/vol] Dubos medium albumin) in roller bottles at 37°C, and bacteria were harvested during the mid-exponential phase of growth. Total RNA was extracted using a Fast RNA Pro Blue kit (Bio 101 Systems). DNA microarray hybridizations were carried out as previously described (23) using a whole-genome microarray of M. tuberculosis H37Rv (version two) prepared in the laboratory of P. Butcher (St. George's, University of London). Fluorescence-labeled cDNA copies of total RNA (5 to 10 mg) were generated by direct incorporation of Cy3- or Cy5-labeled dCTP (Amersham). The two labeled samples to be compared (Cy3 and Cy5) were combined and purified and then hybridized to the microarray under a Lifterslip. Data were obtained from six slides, including dye swaps. Microarray slides were scanned for fluorescence with an Axon 4000A microarray scanner. Grids were fitted to the raw microarray images, and images were quantified using Bluefuse software (BlueGnome Ltd., Cambridge, United Kingdom).

Analysis of DNA microarray data.The array data in the Bluefuse format were analyzed using Limma (version 2.10.4 [26]) in the R statistical language (version 2.5.0). The quality of the spots as determined by the Bluefuse application was used to weight the data. Spots with quality confidence scores below 0.1 (rated E) were excluded from further analysis. The data were normalized within arrays using Loess and were subsequently quantile normalized between arrays. Differentially expressed genes were identified by using the functions lmFit and eBayes. Technical replication of dye swaps was controlled by using the duplicate correlation function in Limma. Multiple testing was applied by using the Benjamini-Hochberg false discovery rate at a P value of <0.05.

Real-time qPCR.Real-time quantitative PCR (qPCR) was performed with a Rotor-Gene 6000 (Corbett Life Science, Sydney, Australia) machine, using SensiMixPlus SYBR (Quantace, London, United Kingdom) in 10-μl-final-volume reaction mixtures. The RNA was from the same samples used for the microarray experiments. One microgram was reverse transcribed using a Quantitect reverse transcription kit (Qiagen). One microliter of the reverse transcription reaction mixture was used per 10 μl of real-time PCR. The level of expression of each gene was established based on a standard curve and was normalized to the expression level of the control gene sigA. The change in induction (n-fold) was determined by using a 2-standard curve relative quantitation for each gene of interest compared to that of sigA, using Rotorgene software. The oligonucleotides used for Rv0099 consisted of the forward primer Myc1709 (5′-CTGTCGGTGCTGGCGTGTGC-3′) and the reverse primer Myc1710 (5′-AGATGCCATCTTGCTCCCTG-3′); for Rv3616c, the forward primer Myc1711 (5′-CCGGGTGATGGCTGGTTAGG-3′) and the reverse primer Myc1712 (5′-AGGCTGATGAGCTGACGAT-3′); and for Rv3839, the forward primer Myc1713 (5′-GCGGTTCCGGTCGATCGTGG-3′) and the reverse primer Myc1714 (5′-CGGATCCACACCAGCGAACG-3′). For the normalizer gene sigA, the forward primer was Myc1707 (5′-CTCGACGCTGAACCAGACCT-3′), and the reverse primer was Myc1708 (5′-AGGTCTTCGTGGTCTTCGT-3′).

Microarray data accession numbers.The raw data from the DNA microarray experiments were deposited in BμG@Sbase under accession number E-BUGS-61 (http://bugs.sgul.ac.uk/E-BUGS-61 ) and in the ArrayExpress database (http://www.ebi.ac.uk/arrayexpress/ ) under accession number E-BUGS-61. The array design is available at BμG@Sbase (accession number A-BUGS-23 [http://bugs.sgul.ac.uk/A-BUGS-23 ]) and also ArrayExpress (accession number A-BUGS-23).

The BCG allele of the CRP results in altered gene expression in M. tuberculosis.To ascertain the effect of SNPs on function, RNA was extracted from exponential phase cultures of M. tuberculosis strains carrying the Rv3676 deletion (Δcrp) and complemented with a plasmid carrying either the M. tuberculosis Rv3676 sequence (pcrp_Mtb) or a mutated version carrying the BCG (pcrp_BCG) sequence, both transcribed from the native Rv3676 promoter. (The BCG sequence was derived by site-directed mutagenesis from M. tuberculosis H37Rv DNA, see Materials and Methods.) This RNA was analyzed by using whole-genome M. tuberculosis microarrays. A total of six hybridizations, including dye swaps, was carried out from three replicate cultures. This experiment identified a number of genes that were altered in expression in the strain carrying pcrp_BCG compared with that in the strain carrying pcrp_Mtb (Table 1). For three genes (Rv0099, Rv3616c, and Rv3839) the microarray data were confirmed, using the same RNA samples, by measuring the amount of mRNA, using real-time qPCR, standardizing to the amount of mRNA of a normalizer gene, sigA, coding for the primary sigma factor. The expression level of sigA did not change in the array analysis under these conditions, and this gene has previously been shown to exhibit constant expression during exponential growth and so can be used as an internal standard for mRNA quantification (11). The results obtained were comparable with the results from the microarray experiment; thus, by qPCR, Rv0099 exhibited a 2.43-fold higher expression when pcrp_BCG rather than pcrp_Mtb was present and 2.15-fold higher expression by microarray. The comparable figures for Rv3616c were 2.0-fold lower expression by qPCR and 2.2-fold lower by microarray and for Rv3839, 2.78-fold lower by qPCR and 4.08-fold lower by microarray.

A number of genes with altered expression were previously identified from a microarray analysis comparing the Rv3676 mutant with the wild-type M. tuberculosis H37Rv (23). These genes included fadD10, ahpC, and lipQ and the operon of Rv3616c-Rv3613c. Of these, the last three genes/operons had recognizable CRP-binding sites in their promoters, and among these there was a consistent pattern of expression compared with the expression in the Rv3676 (Δcrp) deletion mutant. Thus, in the previously published experiments comparing the Δcrp mutant with the wild-type M. tuberculosis H37Rv (23), ahpC exhibited decreased expression in the Δcrp mutant compared to expression in the wild type, and in the present paper, exhibited increased expression when the Δcrp mutant was complemented by pcrp_BCG compared to complementation by pcrp_Mtb. This expression pattern was reversed with Rv3616c and lipQ.

Some genes that were notably altered in their expression when Rv3676 was deleted did not, however, appear at all in the list of differentially expressed genes in the present experiment. These included rpfA and whiB1, both of which have been proven by band shift experiments and transcriptional fusion experiments to be directly controlled by the Rv3676 protein (1, 23; D. Hunt, unpublished data). Moreover, other genes that appeared to be differentially expressed when the BCG allele was present did not appear in the list of genes affected by the deletion of Rv3676. Some of these genes, such as Rv2989 and Rv0250c, had credible potential Rv3676 binding sites in their promoter regions, based on the consensus sequence defined in previous experiments (3, 23). Others, however, notably sigC, did not, and could therefore be the secondary result of alterations in transcription of other genes. Although the sigC gene showed the largest increase compared to that of the wild type when the BCG allele was present, the increase was only 2.3-fold. Perhaps more notable was the decreased expression of Rv3839 (4.08-fold decrease) and Rv3840 (2.42-fold decrease), which comprise a putative operon, since Rv3840 is annotated as a possible transcriptional regulatory protein that could therefore affect the expression of other genes. Like sigC, these genes do not have a very credible CRP-binding site, although Rv3839 does have the invariant GT and AC nucleotides, as identified by Bai et al. (2).

Modeling of the CRP structure: effect of the BCG SNPs.To explain the changes in gene expression, the structure of the CRP encoded by Rv3676/Mb3700 was modeled on the known structure of CRP from E. coli. We have analyzed the effect of the SNPs by reference to this E. coli structure (1J59). The Leu47Pro substitution in the nucleotide-binding domain of CRP from the BCG Danish, Glaxo, and Pasteur strains (equivalent to Leu39 in the E. coli structure) occurs in a buried β-sheet strand with the side chain pointing into the core of the protein. The backbone nitrogen and carbonyl oxygen atoms form hydrogen bonds with the neighboring β-strand residue Ile97 (Fig. 2i). The change from Leu47 to Pro47 in BCG entails the loss of one hydrogen bond (Fig. 2ii) to the neighboring residue Met104, but this is accommodated by the structure (prolines are often found in antiparallel β sheets), and therefore, this SNP is expected to cause no change in function.

• Open in new tab
• Download powerpoint

FIG. 2.

Interactions between amino acid residues in the wild type (M. tuberculosis and M. bovis) and those in the BCG CRP. (i) Leu47 residue of the wild-type CRP binding to residue Ile97; (ii) SNP Pro47 of BCG binding to Met104, showing the loss of one hydrogen bond; (iii) Glu178 in the wild-type CRP shows absence of the hydrogen bond with the phosphate backbone of DNA; (iv) SNP Lys178 in BCG maintains this hydrogen bond with the phosphate backbone of DNA, as occurs in E. coli.

However, the SNP in the DNA binding domain is predicted to have a more significant effect. In the HTH region, Gln170 in E. coli forms a hydrogen bond with the phosphate backbone of the DNA (predominantly with the phosphate oxygen atom of residue G3 in the C chain of the DNA). The model shows that the equivalent residue in BCG, Lys178 (Fig. 2iv), is likely to maintain this hydrogen bond pattern, but Glu178 (Fig. 2iii) in M. tuberculosis/M. bovis would not. Thus, this SNP is likely to enhance the binding of CRP to DNA and could explain the changes in gene expression.

The BCG allele of CRP complements the Rv3676 deletion mutant of M. tuberculosis: effect on growth in vitro and in vivo.The strain of M. tuberculosis (Δcrp) with a deletion of most of the coding region of Rv3676 was previously found to have a growth defect in vitro (23), in infections in macrophages, and by intravenous infection in mice. We have complemented this mutant with a plasmid carrying the M. tuberculosis Rv3676 sequence (pcrp_Mtb) and with the same plasmid carrying the BCG sequence (pcrp_BCG). As expected from our previous work (23), pcrp_Mtb fully complemented the in vitro growth defect of the Rv3676 deletion mutant. However, it was also evident that pcrp_BCG also complemented this growth defect. Therefore, with regard to growth in vitro, the BCG Rv3676 sequence fully complemented the Rv3676 deletion mutant of M. tuberculosis (Fig. 3).

• Open in new tab
• Download powerpoint

FIG. 3.

Growth of the M. tuberculosis H37Rv wild-type strain, the Rv3676 deletion mutant, and the latter complemented with either the M. tuberculosis Rv3676 sequence or the BCG Rv3676 sequence during in vitro liquid culture in rolling bottles at 37°C in Dubos medium. Each strain was grown to an optical density at 600 nm of 0.5 and diluted 10-fold in fresh medium, and samples were taken for optical density readings at 600 nm.

Deletion of Rv3676 in M. tuberculosis resulted in growth defects in macrophages and in mice, which could be complemented by the wild-type M. tuberculosis Rv3676 allele (23). A similar complementation test was performed with strains carrying some of the plasmids mentioned in the previous section. In an infection of unactivated mouse bone marrow-derived macrophages, only the strain with the Rv3676 deletion was attenuated for growth; the Δcrp mutant strains complemented by either pcrp_Mtb or pcrp_BCG grew as well as the wild-type M. tuberculosis H37Rv (Fig. 4).

• Open in new tab
• Download powerpoint

FIG. 4.

Growth of the M. tuberculosis H37Rv wild-type strain, the Rv3676 deletion mutant, and the latter complemented with either the M. tuberculosis Rv3676 sequence or the BCG Rv3676 sequence in bone marrow-derived macrophages. Macrophages were isolated from BALB/c mice as described previously (23) and infected at a multiplicity of infection of one bacterium to two macrophages with each strain. The survival and multiplication of the M. tuberculosis strains were determined by CFU counts. The experiment was performed twice and produced similar results; the results of a representative experiment are shown. The results for each time point are the means of CFU determinations performed for triplicate infections, and the error bars show the standard deviations. The asterisk indicates that the result is significantly different from that of the wild type by the two-tailed Student's t test for groups of unequal variance (P < 0.05), as well as by single-factor analysis of variance (P < 0.01). WT, wild type.

Finally to determine whether the SNPs of Rv3676 in BCG affected the growth of M. tuberculosis in vivo, BALB/c mice were injected intravenously and subsequent growth of the bacteria in the lungs was measured (Fig. 5). As found previously, the Δcrp mutant exhibited greatly decreased growth in the lungs of infected mice. This attenuation was fully complemented by both pcrp_Mtb and pcrp_BCG (Fig. 5).

• Open in new tab
• Download powerpoint

FIG. 5.

Growth of the M. tuberculosis Rv3676 deletion mutant and complemented derivatives in a mouse intravenous infection. BALB/c mice were inoculated intravenously with approximately 5 × 10⁵ CFU of each strain. The survival and multiplication of the M. tuberculosis strains in the lungs were determined by CFU counts and are shown for the wild-type H37Rv strain and the Rv3676 mutant and the mutant complemented with the M. tuberculosis Rv3676 sequence or the BCG sequence. The experiment was performed twice and produced similar results; the results of a representative experiment are shown. The results for each time point are the means of CFU determinations performed for organs from three mice, and the error bars show the standard deviations. The starting inocula were similar for all four strains, as follows: wild type (WT), 2.78 × 10⁵; CRP knockout (KO), 6.25 × 10⁵; WT complement, 6.75 × 10⁵; BCG complement, 9.25 × 10⁵. The asterisk indicates that the result is significantly different from that of the wild type by the two-tailed Student t test for groups of unequal variance (P < 0.05).

Thus, the evidence from the modeling results, showing that the SNPs present in the CRP of BCG were not expected to have a drastic effect on its function, lying as they did outside the amino acid residues critical for nucleotide binding and DNA binding, was corroborated by the phenotypes of the complemented derivatives of the Δcrp mutant. Although the change in the HTH region did indicate an enhanced ability to bind to DNA, resulting in changes in gene expression, these changes were not sufficient to result in growth attenuation.