Proteomic and biological diversity of closely related clinical isolates of Mycobacterium tuberculosis
Date
2009
Journal Title
Journal ISSN
Volume Title
Abstract
Recent studies have indicated that some Mycobacterium tuberculosis {Mtb) strains may be more easily transmitted and successful in causing disease. While this phenotypic diversity exits and is well documented, protein variation between Mtb strains appears to be small. Historically, the analysis of virulence determinants in Mtb has focused on the comparison of virulent to avirulent Mtb strains as well as on the comparison of single gene knock-out mutants to Mtb reference strains. Biological and biochemical comparison of Mtb clinical isolates with different degrees of virulence, although more complex than those described above, might provide additional information regarding the natural phenotypic variability that exists in Mtb. The focus of this study was to determine if a group of closely related clinical isolates of Mtb (BE, C28 and H6) as well as the more distantly related Mtb CDC 1551 present differences in both, virulence in the guinea pig model of TB infection and in the protein expression profiles of secreted and cytosolic fractions. The three closely related Mtb isolates comprise the S75 group, which has been reported to be very successful in causing disease. In contrast, Mtb CDC 1551, although reported to be highly transmissible, is associated with only a few number of active TB cases. To test the hypothesis that closely related Mtb strains would have similar virulence between each other, but higher virulence when compared to Mtb CDC 1551, guinea pigs were infected with each Mtb strain and time-to-death analysis was performed. Proteomes of secreted and cytosolic fractions of these strains were also analyzed and compared by 2D-GE and isobaric tag labeling (iTRAQ) followed by tandem mass spectrometry in order to test whether protein levels of these strains would present statistically significant differences when compared to each other and when compared to Mtb CDC1551.
Unexpectedly, differences in both virulence and protein levels between the closely related Mtb clinical isolates were identified. Strains Mtb C28 and Mtb H6 appeared to be more virulent than Mtb BE based on analysis of survival curves of infected guinea pigs. Given that differences in virulence between these closely related strains were identified, the analysis of secreted and cytosolic proteomes was focused in the identification of differential proteins between Mtb BE and the more virulent strains Mtb C28 and Mtb H6 that could explain the differences observed in the survival of guinea pigs infected with these strains. Analysis of proteomes by 2D-GE allowed the identification of 13 spots presenting statistically significant differences between at least 2 of the four analyzed strains. However, most of the differences were between Mtb CDC1551 and the S75 group, indicating that 2D-GE might not be able to resolve subtle dif differences between closely related Mtb strains. A rigorous statistical analysis of iTRAQ data allowed the identification and quantification of 101 and 139 proteins in the secreted and cytosolic fraction respectively. Similar to 2D-GE, most of the differences were observed between Mtb CDC 1551 and the S75 group. Several differences in protein levels were also observed between closely related strains, especially between Mtb BE and both, Mtb C28 and Mtb H6. Proteins involved in virulence, detoxification and adaptation were more abundant in the more virulent strains C28 and H6. Similarly, enzymes related to metabolic pathways (GltA2, SucC, Gndl, Eno) and proteins involved in nutrient uptake and storage (BfrB, ViuB, TB15.3 and SseC2) were also present in higher quantity in these strains when compared to Mtb BE. In contrast, proteins related to cell wall and cell processes were more abundant in Mtb BE. These observations allowed us to construct hypotheses regarding the higher virulence of Mtb C28 and Mtb H6 in the guinea pig model, including an increased ability for adaptation of these Mtb strains to the challenging host environment, a better utilization of energy sources, as well as an advantage for nutrient uptake. Future studies are proposed and discussed to test the hypotheses and to validate our results in an 'in vivo' model.
In a separate study, and based on the fact that the most remarkable proteomic difference between Mtb CDC 1551 and the other three Mtb clinical isolates was a considerable higher amount of an immunogenic protein called Cfp2, preliminary analysis of this protein secretion mechanisms, with an emphasis on its signal peptide, was performed. In addition, possible Cfp2 protein-protein interactions were identified by pull-down assays followed by Tandem MS/MS. As a first step to understand the processing of this protein for secretion, it was demonstrated that a substitution at position -3 of the cleavage site reduced the secretion of this protein. However, mutations in other positions (-1, +1, +2) did not decrease the secretion of this protein when compared to the wild type. Cfp2 co-eluting proteins were identified, including several proteins involved in fatty acid biosynthesis, indicating that Cfp2 might have a role in this cellular process.
In conclusion, the combined use of the guinea pig model and proteomic analysis demonstrated that closely related strains of Mtb not only presented differences in virulence, but they also presented several differences in the protein expression levels in both cytosol and secreted protein fractions. The pattern of expression of certain groups of proteins suggests a possible relation to the different degrees of virulence of each of the strains and provides a foundation for further validation of these proteins in the context of Mtb virulence.
Preliminary studies on Cfp2 generated protocols for Cfp2 native purification and recombinant production and provided the basis for further studies on Mtb signal peptides as well as the possible role of Cfp2 in fatty acid biosynthesis.
Unexpectedly, differences in both virulence and protein levels between the closely related Mtb clinical isolates were identified. Strains Mtb C28 and Mtb H6 appeared to be more virulent than Mtb BE based on analysis of survival curves of infected guinea pigs. Given that differences in virulence between these closely related strains were identified, the analysis of secreted and cytosolic proteomes was focused in the identification of differential proteins between Mtb BE and the more virulent strains Mtb C28 and Mtb H6 that could explain the differences observed in the survival of guinea pigs infected with these strains. Analysis of proteomes by 2D-GE allowed the identification of 13 spots presenting statistically significant differences between at least 2 of the four analyzed strains. However, most of the differences were between Mtb CDC1551 and the S75 group, indicating that 2D-GE might not be able to resolve subtle dif differences between closely related Mtb strains. A rigorous statistical analysis of iTRAQ data allowed the identification and quantification of 101 and 139 proteins in the secreted and cytosolic fraction respectively. Similar to 2D-GE, most of the differences were observed between Mtb CDC 1551 and the S75 group. Several differences in protein levels were also observed between closely related strains, especially between Mtb BE and both, Mtb C28 and Mtb H6. Proteins involved in virulence, detoxification and adaptation were more abundant in the more virulent strains C28 and H6. Similarly, enzymes related to metabolic pathways (GltA2, SucC, Gndl, Eno) and proteins involved in nutrient uptake and storage (BfrB, ViuB, TB15.3 and SseC2) were also present in higher quantity in these strains when compared to Mtb BE. In contrast, proteins related to cell wall and cell processes were more abundant in Mtb BE. These observations allowed us to construct hypotheses regarding the higher virulence of Mtb C28 and Mtb H6 in the guinea pig model, including an increased ability for adaptation of these Mtb strains to the challenging host environment, a better utilization of energy sources, as well as an advantage for nutrient uptake. Future studies are proposed and discussed to test the hypotheses and to validate our results in an 'in vivo' model.
In a separate study, and based on the fact that the most remarkable proteomic difference between Mtb CDC 1551 and the other three Mtb clinical isolates was a considerable higher amount of an immunogenic protein called Cfp2, preliminary analysis of this protein secretion mechanisms, with an emphasis on its signal peptide, was performed. In addition, possible Cfp2 protein-protein interactions were identified by pull-down assays followed by Tandem MS/MS. As a first step to understand the processing of this protein for secretion, it was demonstrated that a substitution at position -3 of the cleavage site reduced the secretion of this protein. However, mutations in other positions (-1, +1, +2) did not decrease the secretion of this protein when compared to the wild type. Cfp2 co-eluting proteins were identified, including several proteins involved in fatty acid biosynthesis, indicating that Cfp2 might have a role in this cellular process.
In conclusion, the combined use of the guinea pig model and proteomic analysis demonstrated that closely related strains of Mtb not only presented differences in virulence, but they also presented several differences in the protein expression levels in both cytosol and secreted protein fractions. The pattern of expression of certain groups of proteins suggests a possible relation to the different degrees of virulence of each of the strains and provides a foundation for further validation of these proteins in the context of Mtb virulence.
Preliminary studies on Cfp2 generated protocols for Cfp2 native purification and recombinant production and provided the basis for further studies on Mtb signal peptides as well as the possible role of Cfp2 in fatty acid biosynthesis.
Description
Rights Access
Subject
Mycobacterium tuberculosis
proteomics
microbiology
biochemistry