This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hunt, D. M.
Right arrow Articles by Buxton, R. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hunt, D. M.
Right arrow Articles by Buxton, R. S.

 Previous Article  |  Next Article 

Infection and Immunity, May 2008, p. 2227-2234, Vol. 76, No. 5
0019-9567/08/$08.00+0     doi:10.1128/IAI.01410-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

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{triangledown}

Debbie M. Hunt,1,{ddagger} José W. Saldanha,2,{ddagger} John F. Brennan,1 Pearline Benjamin,1 Molly Strom,3 Jeffrey A. Cole,4 Claire L. Spreadbury,5,{dagger} and Roger S. Buxton1*

Division of Mycobacterial Research,1 Division of Mathematical Biology,2 Division of Molecular Neuroendocrinology, MRC National Institute for Medical Research, Mill Hill, London NW7 1AA, United Kingdom,3 School of Biosciences and Division of Infection and Immunity,4 The Medical School, University of Birmingham, Birmingham B15 2TT, United Kingdom5

Received 19 October 2007/ Returned for modification 20 November 2007/ Accepted 26 February 2008


arrow
ABSTRACT
 
Single nucleotide polymorphisms (SNPs) are present in the global transcriptional regulator cyclic AMP (cAMP) receptor protein (CRP) of the attenuated vaccine strain Mycobacterium bovis, bacillus Calmette-Guérin (BCG). We have found that these SNPs resulted in small but significant changes in the expression of a number of genes in M. tuberculosis when a deletion of the Rv3676 CRP was complemented by the BCG allele, compared to complementation by the M. tuberculosis allele. We can explain these changes in gene expression by modeling the structure of the mycobacterial protein on the known structure of CRP from Escherichia coli. Thus, the SNP change in the DNA-binding domain, Lys178, is predicted to form a hydrogen bond with the phosphate backbone of the DNA, as does the equivalent residue in E. coli, whereas Glu178 in M. tuberculosis/M. bovis does not, thus explaining the stronger binding reported for CRP of BCG to CRP-binding sites in mycobacterial DNA. In contrast, the SNP change in the nucleotide binding domain (Leu47Pro) is predicted to result in the loss of one hydrogen bond, which is accommodated by the structure, and would not therefore be expected to cause any change in function relating to cAMP binding. The BCG allele fully complemented the growth defect caused by the deletion of the Rv3676 protein in M. tuberculosis, both in vitro and in macrophage and mouse infections, suggesting that these SNPs do not play any role in the attenuation of BCG. However, they may have allowed BCG to grow better under the in vitro-selective conditions used in its derivation from the M. bovis wild type.


arrow
INTRODUCTION
 
Tuberculosis (TB), caused by the bacterium Mycobacterium tuberculosis, remains the infectious disease that kills the largest number of people worldwide, despite the fact that a vaccine, bacillus Calmette-Guérin (BCG), has been available since the 1920s and is relatively inexpensive. BCG is the world's most widely used vaccine, but it does not provide complete protection against TB. At most, it provides only 80% protection (14), and is particularly effective against disseminated forms of TB in children, such as miliary TB and tuberculosis meningitis. However, the protection conferred against pulmonary TB in adults, the majority of carriers of the disease burden, has been very variable in clinical trials, depending on the population, the country, and the BCG strain used. BCG does not in fact provide adequate levels of protection in Africa, India, and some parts of the United States. So, new vaccines are required, and a great deal of current research is directed toward achieving this end (4), including the use of BCG as the starting point for new vaccines (20). It is therefore important to have as much information as possible about the genomic and virulence status of the BCG strain to know how it differs from virulent strains.

Using the attenuated strain vaccine production techniques pioneered by Louis Pasteur, BCG was derived by Calmette and Guérin from virulent Mycobacterium bovis, the agent of bovine TB and a close relative of M. tuberculosis, by repeated subculturing on potato-glycerin-ox bile agar. Between 1906 and 1919, Calmette and Guérin performed 230 passages and reported a gradual loss of virulence in various animal models (10), so that eventually, it was used as the vaccine to prevent human TB. In the last few years, a considerable amount of work has been undertaken to elucidate the genomic composition of BCG in comparison to its virulent parent strain, M. bovis (for a review, see reference 7), resulting in the description of a number of genomic deletions. Conclusive evidence has accumulated that the deletion of one chromosomal region in particular, region of difference 1 (RD1), contributes greatly to the attenuation of this vaccine strain. Thus, deletion of the RD1 region from the virulent H37Rv strain of M. tuberculosis caused an attenuation similar to that exhibited by M. bovis BCG (19). Of crucial importance, the restoration of RD1 by a gene knock-in resulted in colonial morphology and a virulence level in a mouse infection model closer to that of the virulent strains of M. tuberculosis and M. bovis (22). However, complementation with RD1 did not completely restore virulence levels to those seen with M. tuberculosis or M. bovis (22), leading to speculation that other polymorphisms that have occurred in BCG also contribute to its attenuation. Besides deletions, these polymorphisms could include point mutations (single nucleotide polymorphisms [SNPs]) or small deletions that were not identified using the large scale genomic techniques such as comparative DNA microarray analysis. A number of such point mutation differences have now been found by DNA sequencing (8) that might affect BCG attenuation. This paper examines one of the first BCG polymorphisms to be described in the Mb3700 gene coding for a transcriptional regulator of the cyclic AMP (cAMP) receptor protein (CRP)-fumarate and nitrate reduction regulator (FNR) family (27).

The CRP-FNR family consists of global transcriptional regulator proteins that are all predicted to be structurally related to CRP. The archetypal CRP fold (25) is a versatile structure that has evolved to accommodate different functional specificities in signal perception, DNA binding, and interaction with RNA polymerase to allow different family members to respond to a wide range of signals (13, 17). The best-characterized member of the family is CRP itself from Escherichia coli. This protein controls the expression of numerous genes in response to changes in the intracellular concentration of cAMP (12). Upon binding to cAMP, the cAMP-CRP complex binds to promoters containing DNA sequences related to the consensus TGTGANNNNNNTCACA (5). After binding to promoter DNA, CRP recruits RNA polymerase and specific protein-protein contacts are established that promote the transcription of target genes (reviewed in reference 9), although at some promoters, CRP represses transcription, for example, by promoter occlusion.

Members of this protein family regulate a diversity of physiological functions in many bacteria and respond to a variety of environmental and metabolic signals (13). The protein encoded by Mb3700 (identical in sequence to Rv3676 in M. tuberculosis) is more similar to the CRP protein of E. coli, which binds to cAMP, than to any of the other members of the family (23), and there is some experimental evidence that it does bind to cAMP (1, 3; M. Stapleton and J. Green, unpublished data).

In the present paper we examine the molecular consequence of the M. bovis BCG SNPs in CRP and determine their effect on attenuation.


arrow
MATERIALS AND METHODS
 
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 pcrpMtb) transcribed from its own promoter or mutated versions carrying the BCG Rv3676 (pcrpBCG) 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).


arrow
RESULTS
 
The amino acid sequence of the CRP from M. tuberculosis (Rv3676) (which is identical to the sequence of the Mb3700 gene from M. bovis) is shown in comparison to the sequences of CRP from E. coli, Haemophilus influenzae, Pasteurella multocida, and Salmonella enterica serovar Typhimurium, in Fig. 1. Also shown are the secondary structure assignments, either predicted (see Materials and Methods) or, for the E. coli sequence, taken from the X-ray structure of CRP in complex with DNA at 2.5-Å resolution, as described by Parkinson et al. (21). The sequence of the protein (Fig. 1) from all M. bovis BCG strains (e.g., Russia, Moreau, Birkhaug, Danish, Glaxo, Pasteur, and Tokyo) differs from the wild-type sequence (from M. bovis or M. tuberculosis) in an amino acid residue in the HTH of the DNA-binding domain (Glu178Lys). In addition, the Danish, Glaxo, and Pasteur strains also differ in a residue in the cNMP domain (Leu47Pro) (27). In all experiments described, both of the BCG SNPs, in the HTH and the NMP, are present.


Figure 1
View larger version (38K):
[in this window]
[in a new window]

  		FIG. 1. Multiple protein sequence alignment of Mycobacterium tuberculosis Rv3676 (mtb), Escherichia coli (ecoli), Haemophilus influenzae (haein), Pasteurella multocida (pasmu), and Salmonella enterica serovar Typhimurium (salty) CRPs. Numbering is presented according to the M. tuberculosis sequence. Differences in amino acid sequence between the M. tuberculosis/M. bovis and BCG strains are shown above the alignment, and the exclamation symbols represent sites of point mutagenesis to alanine at positions 188 and 189. Heavy shading represents {alpha} helix and light shading β strand. The secondary structure assignments were predicted using PHD (24), except for the E. coli sequence, for which the assignments were taken from the X-ray structure 1J59 (21), using DSSP software (15). Residue conservation is shown immediately below the sequence alignment. Asterisk, fully conserved; colon, strongly conserved; period, weakly conserved.

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 ({Delta}crp) and complemented with a plasmid carrying either the M. tuberculosis Rv3676 sequence (pcrpMtb) or a mutated version carrying the BCG (pcrpBCG) 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 pcrpBCG compared with that in the strain carrying pcrpMtb (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 pcrpBCG rather than pcrpMtb 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.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Genes with altered expression in the crp (Rv3676) mutant complemented with crpMtb or crpBCG

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 ({Delta}crp) deletion mutant. Thus, in the previously published experiments comparing the {Delta}crp mutant with the wild-type M. tuberculosis H37Rv (23), ahpC exhibited decreased expression in the {Delta}crp mutant compared to expression in the wild type, and in the present paper, exhibited increased expression when the {Delta}crp mutant was complemented by pcrpBCG compared to complementation by pcrpMtb. 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.


Figure 2
View larger version (32K):
[in this window]
[in a new window]

  		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 ({Delta}crp) 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 (pcrpMtb) and with the same plasmid carrying the BCG sequence (pcrpBCG). As expected from our previous work (23), pcrpMtb fully complemented the in vitro growth defect of the Rv3676 deletion mutant. However, it was also evident that pcrpBCG 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).


Figure 3
View larger version (13K):
[in this window]
[in a new window]

  		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 {Delta}crp mutant strains complemented by either pcrpMtb or pcrpBCG grew as well as the wild-type M. tuberculosis H37Rv (Fig. 4).


Figure 4
View larger version (16K):
[in this window]
[in a new window]

  		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 {Delta}crp mutant exhibited greatly decreased growth in the lungs of infected mice. This attenuation was fully complemented by both pcrpMtb and pcrpBCG (Fig. 5).


Figure 5
View larger version (16K):
[in this window]
[in a new window]

  		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 x 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 x 105; CRP knockout (KO), 6.25 x 105; WT complement, 6.75 x 105; BCG complement, 9.25 x 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 {Delta}crp 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.


arrow
DISCUSSION
 
Spreadbury et al. (27) showed that most strains of BCG have two mutations in Mb3700, one in the DNA-binding domain and one in the cNMP-binding domain. Judging from a comparison of different BCG strains (8, 27), the mutation in the DNA-binding region seems to have occurred prior to 1924, whereas the mutation in the cNMP-binding domain occurred later, between 1926 and 1931.

By analyzing the gene expression, using DNA microarrays, we have shown that there was an effect of the BCG SNPs on the expression of a number of genes. Some of the genes with altered expression had credible CRP-binding sites in their promoters, such as lipQ, espA (Rv3616c), Rv0250c, ahpC, and Rv2989. These included some genes where this site experimentally has been shown to be functional, such as espA (Rv3616c) and ahpC (R. Whalan and D. M. Hunt, unpublished data).

By our modeling analysis, we have shown that the BCG SNPs in CRP are predicted to result in minor structural changes, most significantly in the HTH region. Interestingly, in a recent paper, Bai et al. (2) found that the CRPBCG retains a functional interaction with both DNA and cAMP but had a higher binding affinity for CRP-binding sites in mycobacterial DNA than CRPMtb as measured by electrophoretic mobility shift assays and chromatin immunoprecipitation experiments. Our model of the CRP structure can explain this enhanced DNA binding by CRPBCG since the Lys178 SNP in BCG is predicted to form a hydrogen bond with the phosphate backbone of the DNA, but this hydrogen bond is predicted to be absent in M. tuberculosis and M. bovis, although it is present in E. coli. This enhanced binding to DNA could therefore be the cause of the changes in gene expression that we observed. The previously reported (27) decrease in DNA binding of the CRPBCG compared to that of CRPMtb in an E. coli lacZ reporter system is presumably an artifact due to the heterologous nature of this system, since the consensus CRP-binding site in E. coli differs slightly from the site in M. tuberculosis (3, 23). On the other hand, the effect of the Pro47 SNP on the CRP structure was deemed to be minimal and not likely to affect cAMP binding, again consistent with the results reported by Bai et al. (2) that CRPBCG retains a functional association with cAMP. Additionally, there is the possibility that the SNPs alter protein interactions of the CRP, either with RNA polymerase or with other unknown proteins, and that this also could contribute to changes in gene expression.

The increased binding of CRPBCG to DNA compared to that of CRPMtb can explain the pattern of gene expression changes observed in the microarrays among genes with recognizable CRP-DNA binding sites. Thus, ahpC exhibits decreased expression in the CRP (Rv3676) deletion strain, indicating that CRP is activating the expression of this gene, whereas increased DNA binding by CRPBCG in the complemented strain leads to increased ahpC expression. In contrast, in the case of Rv3616c and lipQ, where CRP appears to be acting as a repressor, since expression of these genes is increased in the CRP deletion strain, complementation by CRPBCG results in tighter binding and decreased expression.

Although the SNPs resulted in changes in gene expression, the BCG allele complemented the CRP deletion mutant as well as the wild-type allele in terms of growth in vitro and in vivo, in macrophages and in mice. It appears, therefore, that these SNPs are not a contributing factor to the attenuation of BCG.

The in vitro growth conditions used to isolate BCG from M. bovis are known to have led to the selection of another SNP, causing the reversion of the pykA gene coding for pyruvate kinase (16), which enabled BCG to grow on glycerol as the carbon source. It is possible that the SNPs in CRP also enabled BCG to grow better under the in vitro conditions used by Calmette and Guérin to select the BCG strain.


arrow
ACKNOWLEDGMENTS
 
We thank Douglas Young for critical reading of the manuscript and Jacky Pallas, Bloomsbury Centre for Bioinformatics at University College London, for expert help with the analysis of the microarray data. We acknowledge BµG@S (the Bacterial Microarray Group at St. George's, University of London) and particularly Philip Butcher, Jason Hinds, Kate Gould, and Lucy Brooks for the supply of the M. tuberculosis microarrays and advice, and the Wellcome Trust funding of the multicollaborative microbial pathogen microarray facility under its Functional Genomics Resources Initiative.

The work described in this paper was supported by the Medical Research Council.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Division of Mycobacterial Research, MRC National Institute for Medical Research, Mill Hill, London NW7 1AA, United Kingdom. Phone: 44 020 8816 2225. Fax: 44 020 8906 4477. E-mail: rbuxton{at}nimr.mrc.ac.uk Back

{triangledown} Published ahead of print on 10 March 2008. Back

Editor: R. P. Morrison

{ddagger} The first two authors contributed equally to this work. Back

{dagger} Present address: Centre for Extremophile Research, Department of Biology and Biochemistry, University of Bath, Bath BA2 7AY, United Kingdom. Back


arrow
REFERENCES
 
    1
  1. Agarwal, N., T. R. Raghunand, and W. R. Bishai. 2006. Regulation of the expression of whiB1 in Mycobacterium tuberculosis: role of cAMP receptor protein. Microbiology 152:2749-2756.[Abstract/Free Full Text]
    2. 2
  2. Bai, G., M. A. Gazdik, D. D. Schaak, and K. A. McDonough. 2007. The Mycobacterium bovis BCG cyclic AMP receptor-like protein is a functional DNA binding protein in vitro and in vivo, but its activity differs from that of its M. tuberculosis ortholog, Rv3676. Infect. Immun. 75:5509-5517.[Abstract/Free Full Text]
    3. 3
  3. Bai, G., L. A. McCue, and K. A. McDonough. 2005. Characterization of Mycobacterium tuberculosis Rv3676 (CRPMt), a cyclic AMP receptor protein-like DNA binding protein. J. Bacteriol. 187:7795-7804.[Abstract/Free Full Text]
    4. 4
  4. Baumann, S., A. N. Eddine, and S. H. E. Kaufmann. 2006. Progress in tuberculosis vaccine development. Curr. Opin. Immunol. 18:438-448.[CrossRef][Medline]
    5. 5
  5. Berg, O. G., and P. H. von Hippel. 1988. Selection of DNA binding sites by regulatory proteins. II. The binding specificity of cyclic AMP receptor protein to recognition sites. J. Mol. Biol. 200:709-723.[CrossRef][Medline]
    6. 6
  6. Brooks, B. R., R. E. Bruccoleri, B. D. Olafson, D. J. States, S. Swaminathan, and M. Karplus. 1983. CHARMM, a program for macromolecular energy, minimization, and dynamics calculations. J. Comput. Chem. 4:187-217.[CrossRef]
    7. 7
  7. Brosch, R., and M. A. Behr. 2005. Comparative genomics and evolution of Mycobacterium bovis BCG, p. 155-164. In S. T. Cole, K. D. Eisenach, D. N. McMurray, and W. R. Jacobs, Jr. (ed.), Tuberculosis and the tubercule bacillus. ASM Press, Washington, DC.
    8. 8
  8. Brosch, R., S. V. Gordon, T. Garnier, K. Eiglmeier, W. Frigui, P. Valenti, S. Dos Santos, S. Duthoy, C. Lacroix, C. Garcia-Pelayo, J. K. Inwald, P. Golby, J. N. Garcia, R. G. Hewinson, M. A. Behr, M. A. Quail, C. Churcher, B. G. Barrell, J. Parkhill, and S. T. Cole. 2007. Genome plasticity of BCG and impact on vaccine efficacy. Proc. Natl. Acad. Sci. USA 104:5596-5601.[Abstract/Free Full Text]
    9. 9
  9. Busby, S., and R. H. Ebright. 1999. Transcription activation by catabolite activator protein (CAP). J. Mol. Biol. 293:199-213.[CrossRef][Medline]
    10. 10
  10. Calmette, A., and C. Guérin. 1920. Nouvelles recherches expérimentales sur la vaccination des bovidés contre la tuberculose. Ann. Inst. Pasteur 34:553-560. (In French.)
    11. 11
  11. Dubnau, E., P. Fontan, R. Manganelli, S. Soares-Appel, and I. Smith. 2002. Mycobacterium tuberculosis genes induced during infection of human macrophages. Infect. Immun. 70:2787-2795.[Abstract/Free Full Text]
    12. 12
  12. Gosset, G., Z. G. Zhang, S. N. Nayyar, W. A. Cuevas, and M. H. Saier. 2004. Transcriptome analysis of Crp-dependent catabolite control of gene expression in Escherichia coli. J. Bacteriol. 186:3516-3524.[Abstract/Free Full Text]
    13. 13
  13. Green, J., C. Scott, and J. Guest. 2001. Functional versatility in the CRP-FNR superfamily of transcription factors: FNR and FLP. Adv. Microb. Physiol. 44:1-34.[Medline]
    14. 14
  14. Hart, P. D. 1967. Efficacy and applicability of mass B.C.G. vaccination in tuberculosis control. Br. Med. J. 1:587-592.[Free Full Text]
    15. 15
  15. Kabsch, W., and C. Sander. 1983. Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers 22:2577-2637.[CrossRef][Medline]
    16. 16
  16. Keating, L. A., P. R. Wheeler, H. Mansoor, J. K. Inwald, J. Dale, R. G. Hewinson, and S. V. Gordon. 2005. The pyruvate requirement of some members of the Mycobacterium tuberculosis complex is due to an inactive pyruvate kinase: implications for in vivo growth. Mol. Microbiol. 56:163-174.[CrossRef][Medline]
    17. 17
  17. Korner, H., H. J. Sofia, and W. G. Zumft. 2003. Phylogeny of the bacterial superfamily of Crp-Fnr transcription regulators: exploiting the metabolic spectrum by controlling alternative gene programs. FEMS Microbiol. Rev. 27:559-592.[CrossRef][Medline]
    18. 18
  18. Lee, M. H., L. Pascopella, W. R. Jacobs, Jr., and G. F. Hatfull. 1991. Site-specific integration of mycobacteriophage L5: integration-proficient vectors for Mycobacterium smegmatis, Mycobacterium tuberculosis, and bacille Calmette-Guerin. Proc. Natl. Acad. Sci. USA 88:3111-3115.[Abstract/Free Full Text]
    19. 19
  19. Lewis, K. N., R. Liao, K. M. Guinn, M. J. Hickey, S. Smith, M. A. Behr, and D. R. Sherman. 2003. Deletion of RD1 from Mycobacterium tuberculosis mimics bacille Calmette-Guerin attenuation. J. Infect. Dis. 187:117-123.[CrossRef][Medline]
    20. 20
  20. McShane, H., and A. Hill. 2005. Prime-boost immunisation strategies for tuberculosis. Microbes Infect. 7:962.[CrossRef][Medline]
    21. 21
  21. Parkinson, G., C. Wilson, A. Gunasekera, Y. W. Ebright, R. E. Ebright, and H. M. Berman. 1996. Structure of the CAP-DNA complex at 2.5 Å resolution: a complete picture of the protein-DNA interface. J. Mol. Biol. 260:395-408.[CrossRef][Medline]
    22. 22
  22. Pym, A. S., P. Brodin, R. Brosch, M. Huerre, and S. T. Cole. 2002. Loss of RD1 contributed to the attenuation of the live tuberculosis vaccines Mycobacterium bovis BCG and Mycobacterium microti. Mol. Microbiol. 46:709-717.[CrossRef][Medline]
    23. 23
  23. Rickman, L., C. Scott, D. M. Hunt, T. Hutchinson, M. C. Menéndez, R. Whalan, J. Hinds, M. J. Colston, J. Green, and R. S. Buxton. 2005. A member of the cAMP receptor protein family of transcription regulators in Mycobacterium tuberculosis is required for virulence in mice and controls transcription of the rpfA gene coding for a resuscitation promoting factor. Mol. Microbiol. 56:1274-1286.[CrossRef][Medline]
    24. 24
  24. Rost, B. 1996. PHD: predicting one-dimensional protein structure by profile-based neural networks. Methods Enzymol. 266:525-539.[CrossRef][Medline]
    25. 25
  25. Schultz, S. C., G. C. Shields, and T. A. Steitz. 1991. Crystal structure of a CAP-DNA complex: the DNA is bent by 90 degrees. Science 253:1001-1007.[Abstract/Free Full Text]
    26. 26
  26. Smyth, G. K. 2005. Limma: linear models for microarray data, p. 397-420. In R. Gentleman, V. Carey, S. Dudoit, R. Irizarry, and W. Huber (ed.), Bioinformatics and computational biology solutions using R and Bioconductor. Springer, New York, NY.
    27. 27
  27. Spreadbury, C. L., M. J. Pallen, T. Overton, M. A. Behr, S. Mostowy, S. Spiro, S. J. Busby, and J. A. Cole. 2005. Point mutations in the DNA- and cNMP-binding domains of the homologue of the cAMP receptor protein (CRP) in Mycobacterium bovis BCG: implications for the inactivation of a global regulator and strain attenuation. Microbiology 151:547-556.[Abstract/Free Full Text]
    28. 28
  28. Stover, C. K., V. F. de la Cruz, T. R. Fuerst, J. E. Burlein, L. A. Benson, L. T. Bennett, G. P. Bansal, J. F. Young, M. H. Lee, G. F. Hatfull, et al. 1991. New use of BCG for recombinant vaccines. Nature 351:456-460.[CrossRef][Medline]
    29. 29
  29. Sussman, J. L., D. Lin, J. Jiang, N. O. Manning, J. Prilusky, O. Ritter, and E. E. Abola. 1998. Protein Data Bank (PDB): database of three-dimensional structural information of biological macromolecules. Acta Crystallogr. D Biol. Crystallogr. 54:1078-1084.[CrossRef][Medline]

Infection and Immunity, May 2008, p. 2227-2234, Vol. 76, No. 5
0019-9567/08/$08.00+0     doi:10.1128/IAI.01410-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.



This article has been cited by other articles:

  • Garcia Pelayo, M. C., Uplekar, S., Keniry, A., Mendoza Lopez, P., Garnier, T., Nunez Garcia, J., Boschiroli, L., Zhou, X., Parkhill, J., Smith, N., Hewinson, R. G., Cole, S. T., Gordon, S. V. (2009). A Comprehensive Survey of Single Nucleotide Polymorphisms (SNPs) across Mycobacterium bovis Strains and M. bovis BCG Vaccine Strains Refines the Genealogy and Defines a Minimal Set of SNPs That Separate Virulent M. bovis Strains and M. bovis BCG Strains. Infect. Immun. 77: 2230-2238 [Abstract] [Full Text]  
  • Rampini, S. K., Selchow, P., Keller, C., Ehlers, S., Bottger, E. C., Sander, P. (2008). LspA inactivation in Mycobacterium tuberculosis results in attenuation without affecting phagosome maturation arrest. Microbiology 154: 2991-3001 [Abstract] [Full Text]  

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Hunt, D. M.
Right arrow Articles by Buxton, R. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hunt, D. M.
Right arrow Articles by Buxton, R. S.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»